From 25513156bb113bc391b05a0fe7973a2c1f144b82 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 4 Jul 2026 20:09:16 -0400 Subject: [PATCH 01/41] fix(spawn): isolate docker-compose project names per worktree Two treehouse worktrees of the same project share a leaf directory basename, so Compose's default project-naming collides across them on container/volume/network names. fm-spawn.sh now derives a stable name from the worktree's own pool-slot path and drops it at .treehouse-compose-project (git-excluded, like the other turn-end pointer files), recorded in meta as compose_project=. fm-brief.sh tells ship/scout crewmates to consume it on every docker compose invocation and to avoid fixed host ports as a second collision vector. --- bin/fm-brief.sh | 12 +++++++++ bin/fm-spawn.sh | 70 ++++++++++++++++++++++++++++++++++++++++++++----- 2 files changed, 75 insertions(+), 7 deletions(-) diff --git a/bin/fm-brief.sh b/bin/fm-brief.sh index 15494ca6a..8aec505df 100755 --- a/bin/fm-brief.sh +++ b/bin/fm-brief.sh @@ -235,6 +235,12 @@ This is a SCOUT task: the deliverable is a written report, not a PR. The worktree is your laboratory - install, run, edit, and make scratch commits freely; all of it is discarded at teardown. The report is the only thing that survives, so anything worth keeping must be in it. +This worktree also carries a docker-compose isolation marker at \`.treehouse-compose-project\` in the worktree root. +If this task starts docker or docker-compose, read that file and pass its contents explicitly on every \`docker compose\` invocation, e.g. \`docker compose -p "\$(cat .treehouse-compose-project)" up -d\`. +Exported env vars do not reliably persist between separate tool-call invocations, so if you use \`export COMPOSE_PROJECT_NAME=\$(cat .treehouse-compose-project)\` instead, restate it at the start of every docker-related shell command rather than exporting once and assuming it sticks. +Without this, Compose's default project-naming (derived from the worktree's directory basename) collides with every other worktree of this same project, and both fight over the same container/volume/network names. +A second, related collision vector is fixed host ports: prefer not publishing a hard-coded host port and query the assigned one with \`docker compose port \`, or otherwise pick a deterministic per-worktree-derived port, since two isolated stacks can still fight over a shared host port even with different project names. + # Rules 1. Never push to any remote and never open a PR. 2. Stay inside this worktree; the only files you may write outside it are the report and the status file below. @@ -339,6 +345,12 @@ You are in a disposable git worktree of $REPO, at a detached HEAD on a clean def The path check is authoritative: \`git rev-parse --git-dir\` and \`git rev-parse --git-common-dir\` can help inspect the repo, but they do not prove you are outside the primary checkout. If the top-level path is the primary checkout or not the worktree you were launched in, STOP - do not branch or commit here - append \`blocked: launched in primary checkout, not an isolated worktree\` to the status file and stop. +This worktree also carries a docker-compose isolation marker at \`.treehouse-compose-project\` in the worktree root. +If this task starts docker or docker-compose, read that file and pass its contents explicitly on every \`docker compose\` invocation, e.g. \`docker compose -p "\$(cat .treehouse-compose-project)" up -d\`. +Exported env vars do not reliably persist between separate tool-call invocations, so if you use \`export COMPOSE_PROJECT_NAME=\$(cat .treehouse-compose-project)\` instead, restate it at the start of every docker-related shell command rather than exporting once and assuming it sticks. +Without this, Compose's default project-naming (derived from the worktree's directory basename) collides with every other worktree of this same project, and both fight over the same container/volume/network names. +A second, related collision vector is fixed host ports: prefer not publishing a hard-coded host port and query the assigned one with \`docker compose port \`, or otherwise pick a deterministic per-worktree-derived port, since two isolated stacks can still fight over a shared host port even with different project names. + 1. First action: create your branch: \`git checkout -b fm/$ID\`$SETUP2 # Rules diff --git a/bin/fm-spawn.sh b/bin/fm-spawn.sh index a4ec0890c..4a336c431 100755 --- a/bin/fm-spawn.sh +++ b/bin/fm-spawn.sh @@ -79,6 +79,12 @@ # On success prints: spawned harness= kind= mode= yolo= window= worktree= # mode/yolo are resolved per-project from data/projects.md for ship/scout tasks; # secondmate spawns record mode=secondmate, yolo=off, home=, and projects=. +# Ship/scout spawns also derive a Docker-Compose-safe project name from the +# worktree's own pool-slot path, write it to /.treehouse-compose-project +# (excluded from git via .git/info/exclude, like the per-harness turn-end hook +# files below), and record it in meta as compose_project=, so two +# worktrees of the same project never collide on Compose's default +# project-naming (fm-brief.sh documents how a crewmate consumes it). set -eu SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" @@ -428,6 +434,50 @@ shell_quote() { printf "'" } +# Appends a path (relative to a worktree root) to that worktree's own +# .git/info/exclude, so a firstmate-owned pointer/marker file never blocks +# teardown's dirty-worktree check or leaks into a crewmate's commit. Shared by +# every per-harness turn-end hook file below and the docker-compose isolation +# marker. Operates on $WT, which callers must have already resolved. +exclude_path() { + local rel=$1 EXCL + EXCL=$(git -C "$WT" rev-parse --git-path info/exclude 2>/dev/null || true) + [ -n "$EXCL" ] || return 0 + mkdir -p "$(dirname "$EXCL")" + grep -qxF "$rel" "$EXCL" 2>/dev/null || echo "$rel" >> "$EXCL" +} + +# Derives a unique, Docker-Compose-safe project name from a worktree's own +# pool-slot path (AGENTS.md "Layout and state"; a treehouse pool worktree is +# .../-//), so two worktrees of the same +# project - which always share the same leaf directory basename, e.g. +# "wealthsync" - never collide on Compose's default project-naming (derived +# from cwd basename) for container/volume/network names. Stable across +# respawns into the same pool slot, so a stale stack left behind by a prior +# task in that slot is superseded rather than orphaned under a new name; +# distinct across the fleet because no two live tasks occupy the same slot at +# once. Falls back to the worktree's own leaf name, then the task id, when the +# path does not have two parent directories (e.g. a non-treehouse backend). +fm_compose_project_name() { # + local wt=$1 fallback=$2 slot pool raw name + slot=$(basename "$(dirname "$wt")" 2>/dev/null || true) + pool=$(basename "$(dirname "$(dirname "$wt")")" 2>/dev/null || true) + if [ -n "$slot" ] && [ "$slot" != / ] && [ -n "$pool" ] && [ "$pool" != / ]; then + raw="$pool-$slot" + else + raw=$(basename "$wt") + fi + name=$(printf '%s' "$raw" | tr '[:upper:]' '[:lower:]' | sed 's/[^a-z0-9-]/-/g' | tr -s '-' | sed 's/^-//; s/-$//') + if [ -z "$name" ]; then + name=$(printf '%s' "wt-$fallback" | tr '[:upper:]' '[:lower:]' | sed 's/[^a-z0-9-]/-/g' | tr -s '-' | sed 's/^-//; s/-$//') + fi + case "$name" in + [a-z0-9]*) : ;; + *) name="wt-$name" ;; + esac + printf '%s\n' "$name" +} + model_flag_for_harness() { local harness=$1 model=$2 [ -n "$model" ] && [ "$model" != default ] || return 0 @@ -861,6 +911,18 @@ if [ "$KIND" != secondmate ] && [ "$BACKEND" != orca ]; then validate_spawn_worktree "treehouse get" "$T" fi +# Docker Compose isolation marker (ship/scout only; a secondmate operates in +# its own firstmate home, not a per-task pool worktree). Two treehouse +# worktrees of the same project share a leaf directory basename, so Compose's +# default project-naming collides across them; drop a stable, worktree-scoped +# name here that the crewmate's brief (fm-brief.sh) tells it to pass on every +# docker/docker-compose invocation. +if [ "$KIND" != secondmate ]; then + COMPOSE_PROJECT=$(fm_compose_project_name "$WT" "$ID") + printf '%s\n' "$COMPOSE_PROJECT" > "$WT/.treehouse-compose-project" + exclude_path '.treehouse-compose-project' +fi + # Per-task temp root: /tmp/fm-/ with Go's build temp nested at gotmp/. Go won't # create GOTMPDIR, so mkdir before it is used; fm-teardown removes the whole root. # Nested (not a bare /tmp/fm-/gotmp) so other per-task temp can live alongside @@ -875,13 +937,6 @@ mkdir -p "$TASK_TMP/gotmp" mkdir -p "$STATE" STATE_REAL=$(cd "$STATE" && pwd -P) TURNEND="$STATE_REAL/$ID.turn-ended" -exclude_path() { - local rel=$1 EXCL - EXCL=$(git -C "$WT" rev-parse --git-path info/exclude 2>/dev/null || true) - [ -n "$EXCL" ] || return 0 - mkdir -p "$(dirname "$EXCL")" - grep -qxF "$rel" "$EXCL" 2>/dev/null || echo "$rel" >> "$EXCL" -} if [ "$KIND" != secondmate ]; then case "$HARNESS" in claude*) @@ -1001,6 +1056,7 @@ META_WINDOW=$T echo "tasktmp=$TASK_TMP" echo "model=${MODEL:-default}" echo "effort=${EFFORT:-default}" + [ "$KIND" = secondmate ] || echo "compose_project=$COMPOSE_PROJECT" # backend= is written only for a non-default (non-tmux) backend, so the # default path's meta stays byte-identical (absent backend= means tmux; # data/fm-backend-design-d7's P1 compatibility contract). From 6fea814bb9c3f607e3d97f9a4b728a3bac335088 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 4 Jul 2026 20:54:04 -0400 Subject: [PATCH 02/41] no-mistakes(test): waiting for background test run to complete --- bin/backends/cmux.sh | 10 +++++++--- bin/fm-teardown.sh | 30 ++++++++++++++++++------------ tests/fm-session-start.test.sh | 13 +++++++++---- 3 files changed, 34 insertions(+), 19 deletions(-) diff --git a/bin/backends/cmux.sh b/bin/backends/cmux.sh index 69dc0b53b..c37f003f4 100644 --- a/bin/backends/cmux.sh +++ b/bin/backends/cmux.sh @@ -398,9 +398,13 @@ fm_backend_cmux_parse_target() { # # (fm_backend_zellij_pane_exists) rather than the design sketch's original # read-screen-based suggestion. fm_backend_cmux_surface_exists() { # - local wsid=$1 sfid=$2 - fm_backend_cmux_cli list-panes --workspace "$wsid" --json --id-format uuids 2>/dev/null \ - | jq -e --arg s "$sfid" '[.panes[]? | select(.surface_ids // [] | index($s))] | length > 0' >/dev/null 2>&1 + local wsid=$1 sfid=$2 raw + # Capture the CLI call's own exit status before piping to jq: on an empty + # stdin (the CLI call failed and printed nothing) `jq -e` still exits 0, so + # piping unconditionally would silently treat a failed call as "not found" + # rather than propagating the failure. + raw=$(fm_backend_cmux_cli list-panes --workspace "$wsid" --json --id-format uuids 2>/dev/null) || return 1 + printf '%s' "$raw" | jq -e --arg s "$sfid" '[.panes[]? | select(.surface_ids // [] | index($s))] | length > 0' >/dev/null 2>&1 } # fm_backend_cmux_target_ready: parse the target and verify it is live via diff --git a/bin/fm-teardown.sh b/bin/fm-teardown.sh index a849579b3..a7ec81570 100755 --- a/bin/fm-teardown.sh +++ b/bin/fm-teardown.sh @@ -275,14 +275,17 @@ pr_is_merged() { } # Is the branch's content already present in the up-to-date default branch? Fetches -# first, then 3-way merges the default branch with HEAD: when HEAD introduces nothing -# the default branch does not already contain (e.g. its change landed via squash) the -# merged tree equals the default branch's tree. This isolates branch-only changes, so -# unrelated commits the default branch gained past the merge-base do not count as -# "added". Returns non-zero when inconclusive (no default ref, or a merge conflict), -# so the caller refuses rather than guesses. +# first, then checks every path HEAD changed since its merge-base with the default +# branch: if the default branch already matches HEAD at each of those paths, HEAD +# introduces nothing the default branch does not already contain (e.g. its change +# landed via squash). This isolates branch-only changes, so unrelated commits the +# default branch gained past the merge-base do not count as "added" - equivalent to +# a 3-way merge's resulting tree matching the default branch's tree, but built from +# plain diff/merge-base so it works on git older than 2.38 (which is what `git +# merge-tree --write-tree` requires). Returns non-zero when inconclusive (no default +# ref, or a real difference), so the caller refuses rather than guesses. content_in_default() { - local name ref default_tree merged_tree + local name ref base changed_file name=$(default_branch) || return 1 if git -C "$WT" remote get-url origin >/dev/null 2>&1; then git -C "$WT" fetch --quiet origin "+refs/heads/$name:refs/remotes/origin/$name" >/dev/null 2>&1 || return 1 @@ -292,11 +295,14 @@ content_in_default() { else return 1 fi - default_tree=$(git -C "$WT" rev-parse --quiet --verify "$ref^{tree}" 2>/dev/null) || return 1 - [ -n "$default_tree" ] || return 1 - merged_tree=$(git -C "$WT" merge-tree --write-tree "$ref" HEAD 2>/dev/null) || return 1 - merged_tree=$(printf '%s\n' "$merged_tree" | head -1) - [ "$merged_tree" = "$default_tree" ] + git -C "$WT" rev-parse --quiet --verify "$ref^{commit}" >/dev/null 2>&1 || return 1 + base=$(git -C "$WT" merge-base "$ref" HEAD 2>/dev/null) || return 1 + local -a changed=() + while IFS= read -r -d '' changed_file; do + changed+=("$changed_file") + done < <(git -C "$WT" diff --name-only -z "$base" HEAD -- 2>/dev/null) + [ "${#changed[@]}" -gt 0 ] || return 0 + git -C "$WT" diff --quiet "$ref" HEAD -- "${changed[@]}" 2>/dev/null } # Has the worktree's committed work actually LANDED, though its commits are not diff --git a/tests/fm-session-start.test.sh b/tests/fm-session-start.test.sh index 6d3b0d970..86aeda132 100755 --- a/tests/fm-session-start.test.sh +++ b/tests/fm-session-start.test.sh @@ -350,7 +350,10 @@ EOF make_fake_toolchain "$fakebin" make_fake_ps_claude "$fakebin" # Force a MISSING diagnostic line so the bootstrap section is non-trivial. - rm -f "$fakebin/node" + # Uses chrome-devtools-axi rather than node/tmux/gh: those are common system + # packages that can shadow the fakebin removal via $BASE_PATH's real system + # directories, making the diagnostic flakily disappear depending on the host. + rm -f "$fakebin/chrome-devtools-axi" printf 'window=fm-sess:w1\nkind=ship\n' > "$home/state/task-a.meta" @@ -373,7 +376,7 @@ EOF [ "$context_line" -lt "$fleet_line" ] || fail "CONTEXT did not precede FLEET STATE" [ "$fleet_line" -lt "$next_line" ] || fail "FLEET STATE did not precede NEXT STEP" - missing_line=$(printf '%s\n' "$out" | grep -n 'MISSING: node' | head -1 | cut -d: -f1) + missing_line=$(printf '%s\n' "$out" | grep -n 'MISSING: chrome-devtools-axi' | head -1 | cut -d: -f1) [ -n "$missing_line" ] || fail "MISSING diagnostic did not appear at all" [ "$missing_line" -lt "$fleet_line" ] || fail "actionable MISSING diagnostic was buried after the bulk fleet-state digest" @@ -533,7 +536,9 @@ $rec EOF make_fake_toolchain "$fakebin" make_fake_ps_claude "$fakebin" - rm -f "$fakebin/node" + # See test_output_ordering_diagnostics_lead: chrome-devtools-axi rather than + # node avoids a real system package shadowing the forced-missing tool. + rm -f "$fakebin/chrome-devtools-axi" append_wake "$home/state" signal task-z "needs-decision: pick a library" @@ -542,7 +547,7 @@ EOF # fm-lock.sh's own exact success text. assert_contains "$out" "lock acquired: harness pid" "fm-lock.sh's real output did not appear (composition, not reimplementation)" # fm-bootstrap.sh's own exact MISSING-tool line format. - assert_contains "$out" "MISSING: node (install:" "fm-bootstrap.sh's real detect line did not appear verbatim" + assert_contains "$out" "MISSING: chrome-devtools-axi (install:" "fm-bootstrap.sh's real detect line did not appear verbatim" # fm-wake-drain.sh's real drained record (raw tab-separated queue line). assert_contains "$out" "$(printf 'signal\ttask-z\tneeds-decision: pick a library')" "fm-wake-drain.sh's real drained record did not appear" From e142ad98920bee7dfdaf554f33a96d0fe1628be3 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 4 Jul 2026 21:28:33 -0400 Subject: [PATCH 03/41] no-mistakes(test): fix(tests): make herdr fake CLI counter race-safe against backgrounded server call --- tests/fm-backend-herdr.test.sh | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/tests/fm-backend-herdr.test.sh b/tests/fm-backend-herdr.test.sh index 5186cca2c..f3100d484 100755 --- a/tests/fm-backend-herdr.test.sh +++ b/tests/fm-backend-herdr.test.sh @@ -34,7 +34,6 @@ set -u LOG="${FM_HERDR_LOG:?}" RESP="${FM_HERDR_RESPONSES:?}" COUNT_FILE="$RESP/.count" -next=$(( $(cat "$COUNT_FILE" 2>/dev/null || echo 0) + 1 )) { printf 'HERDR_SESSION=%s' "${HERDR_SESSION:-}" for a in "$@"; do printf '\x1f%s' "$a"; done @@ -44,6 +43,16 @@ if [ "${1:-}" = status ] && [ "${2:-}" = --json ] && [ "${FM_HERDR_SCRIPT_STATUS printf '{"client":{"version":"0.7.1","protocol":14},"server":{"running":true}}\n' exit 0 fi +# fm_backend_herdr_server_ensure launches `herdr server` fire-and-forget +# (backgrounded) so its poll loop can proceed concurrently. No test scripts a +# canned response for it, so it must never read/write the shared call-ordinal +# COUNT_FILE below - doing so races with whichever synchronous call the poll +# loop issues at the same moment, losing an update and permanently shifting +# every later call's response number. +if [ "${1:-}" = server ]; then + exit 0 +fi +next=$(( $(cat "$COUNT_FILE" 2>/dev/null || echo 0) + 1 )) n=$next echo "$n" > "$COUNT_FILE" if [ -f "$RESP/$n.exit" ]; then From edc8313a98c09e545ae8b25e54bc3ef0d9c0f6dd Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 4 Jul 2026 21:47:34 -0400 Subject: [PATCH 04/41] no-mistakes(document): docs: document per-worktree docker-compose project isolation --- docs/architecture.md | 1 + docs/configuration.md | 9 +++++++++ docs/orca-backend.md | 2 +- docs/scripts.md | 4 ++-- 4 files changed, 13 insertions(+), 3 deletions(-) diff --git a/docs/architecture.md b/docs/architecture.md index f2c590f9c..4e71a6a41 100644 --- a/docs/architecture.md +++ b/docs/architecture.md @@ -95,6 +95,7 @@ Codex App support is recorded in `docs/codex-app-backend.md`; it is not selectab Crewmates never intentionally touch your project clone; [treehouse](https://github.com/kunchenguid/treehouse) pools clean worktrees for tmux, herdr, zellij, and cmux tasks, while Orca creates its own worktrees for `backend=orca`. For ship and scout work, `fm-spawn.sh` refuses to launch unless the resolved task path is a real git worktree root that is distinct from the project primary checkout. +Pooled worktrees of the same project share a leaf directory basename, which collides on Docker Compose's default project-naming; see `docs/configuration.md` "Docker Compose project isolation" for the per-worktree `compose_project=` name `fm-spawn.sh` derives to avoid it. The firstmate repo has one extra exposure because it can dispatch crewmates to work on itself. Its operating checkout (`FM_ROOT`) and the disposable crewmate worktrees are all linked git worktrees of the same repository, so the valid discriminator is branch state, not whether the checkout is linked. diff --git a/docs/configuration.md b/docs/configuration.md index d289c4c15..0d45d09d6 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -80,6 +80,15 @@ The caller-facing label remains `fm-`, but the actual cmux workspace title i Test cleanup must use the guarded path described in [`docs/cmux-backend.md`](cmux-backend.md)'s "Test safety" section, never enumerate-and-close every workspace. The `config/backend` file is not inherited by secondmate homes. +## Docker Compose project isolation (per-worktree) + +Two treehouse worktrees of the same project always share the same leaf directory basename (e.g. `wealthsync`), so Docker Compose's default project-naming, derived from the current directory's basename, collides across them on container, volume, and network names. +For every ship/scout spawn, regardless of runtime backend, `fm-spawn.sh` derives a stable, Compose-safe project name from the worktree's own pool-slot path once it is resolved (a treehouse pool worktree is `.../-//`; the name is `-`, lowercased, with non-`[a-z0-9-]` characters replaced by `-` and repeats collapsed), falling back to the worktree's own leaf name or the task id when the path lacks that pool-slot shape (e.g. a non-treehouse backend). +The name is written to `.treehouse-compose-project` in the worktree root, excluded from git via that worktree's own `.git/info/exclude` so it never blocks teardown's dirty-worktree check or leaks into a crewmate's commit, and recorded in `state/.meta` as `compose_project=`. +It is stable across respawns into the same pool slot, not per task id, so a stale Compose stack left by a prior task in that slot is naturally superseded rather than accumulating orphaned volumes under a new name every respawn. +Secondmates never get this marker or meta field; they run in their own firstmate home, not a per-task pool worktree. +`fm-brief.sh` tells the crewmate, in both ship and scout briefs, to read `.treehouse-compose-project` and pass it explicitly on every `docker compose` invocation (e.g. `docker compose -p "$(cat .treehouse-compose-project)" up -d`), since exported env vars do not reliably persist across separate tool-call invocations; it also flags fixed host ports as a related collision vector and recommends `docker compose port ` or another deterministic per-worktree-derived port instead of a hard-coded one. + ## Away-mode supervisor backend (FM_SUPERVISOR_BACKEND / FM_SUPERVISOR_TARGET) The `/afk` sub-supervisor injects escalation digests into firstmate's own pane independently of where new task endpoints are spawned. diff --git a/docs/orca-backend.md b/docs/orca-backend.md index f9ff8b323..22f123777 100644 --- a/docs/orca-backend.md +++ b/docs/orca-backend.md @@ -71,7 +71,7 @@ Spawn: 1. Ensure the project repo is registered in Orca, adding it with `orca repo add --path` when needed. 2. Create an independent Orca worktree with `orca worktree create --repo id: --name fm- --no-parent --setup skip`. 3. Reuse the terminal returned by Orca worktree creation only when it appears in the verified `result.terminal.handle` shape, or create a titled terminal in that worktree when Orca returns only the worktree. -4. Install firstmate's per-harness turn-end hooks in the Orca worktree. +4. Write the docker-compose isolation marker (`.treehouse-compose-project`) and install firstmate's per-harness turn-end hooks in the Orca worktree. 5. Write metadata, then send `GOTMPDIR` export and the selected harness launch through the recorded Orca terminal. Operation routing: diff --git a/docs/scripts.md b/docs/scripts.md index db420026e..1d8fc5810 100644 --- a/docs/scripts.md +++ b/docs/scripts.md @@ -15,7 +15,7 @@ The shared no-mistakes gate refusal used by `fm-spawn.sh`, `fm-send.sh`, and `fm | `fm-bearings-snapshot.sh` | Project the fleet snapshot to the compact TOON bearings view; local-only unless `--include-prs` | | `fm-update.sh` | Fast-forward-only self-update of firstmate and secondmate homes from origin | | `fm-backlog-handoff.sh` | Validate and delegate queued backlog-item moves into a secondmate home | -| `fm-brief.sh` | Scaffold ship, scout, secondmate-charter, and Herdr-lab briefs | +| `fm-brief.sh` | Scaffold ship, scout, secondmate-charter, and Herdr-lab briefs, including the docker-compose isolation note on ship/scout briefs | | `fm-herdr-lab.sh` | Provision and guardedly operate an isolated, never-default Herdr lab session | | `fm-ensure-agents-md.sh` | Ensure a project's real `AGENTS.md`, its `CLAUDE.md` symlink, and the canonical self-governance section | | `fm-guard.sh` | Warn on primary-checkout tangles, pending queued wakes, and stale watcher liveness | @@ -25,7 +25,7 @@ The shared no-mistakes gate refusal used by `fm-spawn.sh`, `fm-send.sh`, and `fm | `fm-arm-command-policy.mjs` | Semantic owner of the watcher-arm PreToolUse policy (docs/arm-pretool-check.md) | | `fm-supervision-instructions.sh` | Render the session-start primary-harness supervision block or the one-line repair instruction | | `fm-home-seed.sh` | Transactionally provision a secondmate home and maintain `data/secondmates.md` | -| `fm-spawn.sh` | Spawn crewmates, scouts, `id=repo` batches, and secondmates on the resolved harness and runtime backend | +| `fm-spawn.sh` | Spawn crewmates, scouts, `id=repo` batches, and secondmates on the resolved harness and runtime backend; ship/scout spawns also derive and record a per-worktree docker-compose project name (`compose_project=`; docs/configuration.md "Docker Compose project isolation") | | `fm-dispatch-select.sh` | Resolve a matched crew-dispatch rule to one concrete profile, owning `quota-balanced` selection | | `fm-backend.sh` | Runtime-backend selection, meta helpers, selector resolution, and operation dispatch | | `fm-backend-hometag-lib.sh` | Shared per-installation home-tag derivation for zellij tab and cmux workspace titles | From cd343d9fe9905e28945ccc6fb97d3857cf40fb56 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 4 Jul 2026 22:40:52 -0400 Subject: [PATCH 05/41] no-mistakes(document): docs already fully synced; added missing docker-compose note to fm-brief.sh header comment --- bin/fm-brief.sh | 3 + tests/fm-secondmate-lifecycle-e2e.test.sh | 4 + tests/fm-spawn-compose-isolation.test.sh | 137 ++++++++++++++++++++++ 3 files changed, 144 insertions(+) create mode 100755 tests/fm-spawn-compose-isolation.test.sh diff --git a/bin/fm-brief.sh b/bin/fm-brief.sh index 8aec505df..c3ffab689 100755 --- a/bin/fm-brief.sh +++ b/bin/fm-brief.sh @@ -34,6 +34,9 @@ # local-only implement on branch, stop and report "ready in branch" (no push/PR); # captain approves, firstmate merges to local main # Ship briefs begin with a worktree-isolation assertion before the branch step. +# Both ship and scout briefs include a docker-compose isolation note telling the +# crewmate to read .treehouse-compose-project and pass it on every docker +# compose invocation (docs/configuration.md "Docker Compose project isolation"). # Scout tasks ignore mode - their deliverable is a report, not a merge. # Every scaffold's status protocol distinguishes the configured # declared-external-wait verb (FM_CLASSIFY_PAUSED_VERB, default "paused") from diff --git a/tests/fm-secondmate-lifecycle-e2e.test.sh b/tests/fm-secondmate-lifecycle-e2e.test.sh index fd71fdb35..099dd2682 100755 --- a/tests/fm-secondmate-lifecycle-e2e.test.sh +++ b/tests/fm-secondmate-lifecycle-e2e.test.sh @@ -130,6 +130,10 @@ phase_spawn() { assert_no_grep 'notify=' "$LOG" "secondmate codex launch included the parent turn-end notify hook" assert_no_grep 'turn-ended' "$LOG" "secondmate codex launch referenced a parent turn-ended signal" assert_no_grep 'treehouse get' "$LOG" "secondmate spawn ran a project treehouse get" + # A secondmate operates in its own persistent home, not a per-task pool + # worktree, so it never gets the docker-compose isolation marker. + assert_absent "$SUB/.treehouse-compose-project" "secondmate spawn wrongly wrote a compose-project marker into its home" + assert_no_grep 'compose_project=' "$meta" "secondmate spawn meta wrongly recorded compose_project=" pass "spawn: launches in the subhome with persistent charter, records routing meta" } diff --git a/tests/fm-spawn-compose-isolation.test.sh b/tests/fm-spawn-compose-isolation.test.sh new file mode 100755 index 000000000..8f9a663d2 --- /dev/null +++ b/tests/fm-spawn-compose-isolation.test.sh @@ -0,0 +1,137 @@ +#!/usr/bin/env bash +# tests/fm-spawn-compose-isolation.test.sh - behavior tests for the +# docker-compose project-name isolation marker (AGENTS.md "Layout and state", +# bin/fm-spawn.sh's fm_compose_project_name). +# +# Two treehouse worktrees of the SAME project always share the same leaf +# directory basename (e.g. both are checked out at ...///wealthsync), +# so Compose's default project-naming (derived from cwd basename) collides +# across them on container/volume/network names. fm-spawn.sh now derives a +# stable name from the worktree's own pool-slot path and drops it at +# /.treehouse-compose-project, git-excluded so it never pollutes a +# crewmate's dirty-worktree check or commits, and records it in the task's +# meta as compose_project=. +# +# Drives the REAL bin/fm-spawn.sh end to end, exactly like +# tests/fm-tangle-guard.test.sh's isolation-abort case: a fake tmux reports +# FM_FAKE_PANE_PATH as the post-`treehouse get` pane cwd, and a real git +# worktree is added at a path shaped like a genuine treehouse pool slot +# (.../-//) so fm_compose_project_name sees +# realistic input. `treehouse` itself is stubbed to exit 0 (unused: the fake +# tmux answers the cwd query directly). +set -u + +# shellcheck source=tests/lib.sh +. "$(dirname "${BASH_SOURCE[0]}")/lib.sh" + +fm_git_identity fmtest fmtest@example.invalid + +# A fake tmux that reports FM_FAKE_PANE_PATH as the pane's cwd after +# `treehouse get`, names the session on '#S', and swallows window ops. +make_spawn_fakebin() { + local dir=$1 fakebin + fakebin=$(fm_fakebin "$dir") + cat > "$fakebin/tmux" <<'SH' +#!/usr/bin/env bash +set -u +case "$*" in + *"#{pane_current_path}"*) printf '%s\n' "${FM_FAKE_PANE_PATH:-}"; exit 0 ;; +esac +case "${1:-}" in + display-message) printf 'firstmate\n'; exit 0 ;; + list-windows) exit 0 ;; + has-session|new-session|new-window|send-keys) exit 0 ;; +esac +exit 0 +SH + chmod +x "$fakebin/tmux" + fm_fake_exit0 "$fakebin" treehouse + printf '%s\n' "$fakebin" +} + +run_spawn() { # [extra-args...] + local home=$1 id=$2 proj=$3 pane=$4 fakebin=$5 + shift 5 + mkdir -p "$home/data/$id" + printf 'brief\n' > "$home/data/$id/brief.md" + FM_ROOT_OVERRIDE='' FM_HOME="$home" \ + FM_STATE_OVERRIDE="$home/state" FM_DATA_OVERRIDE="$home/data" \ + FM_PROJECTS_OVERRIDE="$home/projects" FM_CONFIG_OVERRIDE="$home/config" \ + FM_SPAWN_NO_GUARD=1 FM_FAKE_PANE_PATH="$pane" TMUX="fake,1,0" \ + PATH="$fakebin:$PATH" \ + "$ROOT/bin/fm-spawn.sh" "$id" "$proj" codex "$@" 2>&1 +} + +# --- two worktrees of the same project, same leaf basename, different pool +# slots: this is the exact near-miss from the wealthsync incident this fix +# targets. Each must get its own compose_project name. -------------------- + +test_two_worktrees_same_leaf_get_distinct_compose_names() { + local tmp proj fakebin wt_a wt_b out_a out_b + tmp=$(fm_test_tmproot fm-spawn-compose) + proj=$(fm_git_init_commit "$tmp/proj" && printf '%s\n' "$tmp/proj") + fakebin=$(make_spawn_fakebin "$tmp/fake") + + # Two pool-slot worktrees of the SAME project, sharing the leaf name + # "wealthsync" - mirrors a real treehouse pool: -//. + wt_a="$tmp/wealthsync-323ec0/2/wealthsync" + wt_b="$tmp/wealthsync-323ec0/7/wealthsync" + git -C "$proj" worktree add -q --detach "$wt_a" || fail "setup: worktree add (slot 2) failed" + git -C "$proj" worktree add -q --detach "$wt_b" || fail "setup: worktree add (slot 7) failed" + + out_a=$(run_spawn "$tmp/home" wealthsync-task-a "$proj" "$wt_a" "$fakebin") + expect_code 0 "$?" "spawn into pool slot 2 should succeed" + assert_contains "$out_a" "spawned wealthsync-task-a" "slot-2 spawn did not report success" + + out_b=$(run_spawn "$tmp/home" wealthsync-task-b "$proj" "$wt_b" "$fakebin") + expect_code 0 "$?" "spawn into pool slot 7 should succeed" + assert_contains "$out_b" "spawned wealthsync-task-b" "slot-7 spawn did not report success" + + assert_present "$wt_a/.treehouse-compose-project" "slot-2 worktree missing the compose-project marker" + assert_present "$wt_b/.treehouse-compose-project" "slot-7 worktree missing the compose-project marker" + + local name_a name_b + name_a=$(cat "$wt_a/.treehouse-compose-project") + name_b=$(cat "$wt_b/.treehouse-compose-project") + + [ "$name_a" = "wealthsync-323ec0-2" ] || fail "slot-2 compose project name: expected 'wealthsync-323ec0-2', got '$name_a'" + [ "$name_b" = "wealthsync-323ec0-7" ] || fail "slot-7 compose project name: expected 'wealthsync-323ec0-7', got '$name_b'" + [ "$name_a" != "$name_b" ] || fail "two worktrees of the same project must not collide on the same compose project name" + + assert_grep "compose_project=$name_a" "$tmp/home/state/wealthsync-task-a.meta" "task-a meta missing compose_project=" + assert_grep "compose_project=$name_b" "$tmp/home/state/wealthsync-task-b.meta" "task-b meta missing compose_project=" + + pass "fm-spawn: two worktrees of the same project (same leaf basename, different pool slots) get distinct, stable compose_project names" +} + +# --- the marker must be invisible to the worktree's own git status, so it +# never trips teardown's dirty-worktree check or leaks into a crewmate commit. + +test_marker_is_git_excluded() { + local tmp proj fakebin wt out status_out + tmp=$(fm_test_tmproot fm-spawn-compose-excl) + proj=$(fm_git_init_commit "$tmp/proj" && printf '%s\n' "$tmp/proj") + fakebin=$(make_spawn_fakebin "$tmp/fake") + + wt="$tmp/excltest-9ab1/3/excltest" + git -C "$proj" worktree add -q --detach "$wt" || fail "setup: worktree add failed" + + out=$(run_spawn "$tmp/home" excl-task-c "$proj" "$wt" "$fakebin") + expect_code 0 "$?" "spawn should succeed" + assert_present "$wt/.treehouse-compose-project" "worktree missing the compose-project marker" + + status_out=$(git -C "$wt" status --porcelain) + assert_not_contains "$status_out" ".treehouse-compose-project" "compose-project marker must be git-excluded, not shown as untracked" + + pass "fm-spawn: the compose-project marker is excluded from the worktree's own git status" +} + +# Secondmate spawns never write the marker (a secondmate runs in its own +# persistent firstmate home, not a per-task pool worktree). Setting up a real +# secondmate home's full validation contract (SUB_HOME_MARKER, AGENTS.md, +# bin/, registry) duplicates tests/fm-secondmate-lifecycle-e2e.test.sh's own +# fixture, so that suite's phase_spawn carries the corresponding assertions +# instead of re-deriving the fixture here. + +test_two_worktrees_same_leaf_get_distinct_compose_names +test_marker_is_git_excluded From cbe8d043cf644755e0f1dde85d5b2be4b6abf1c4 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 4 Jul 2026 22:59:32 -0400 Subject: [PATCH 06/41] no-mistakes(lint): Fix unused $out variable flagged by shellcheck in compose isolation test --- tests/fm-spawn-compose-isolation.test.sh | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/fm-spawn-compose-isolation.test.sh b/tests/fm-spawn-compose-isolation.test.sh index 8f9a663d2..496c998ea 100755 --- a/tests/fm-spawn-compose-isolation.test.sh +++ b/tests/fm-spawn-compose-isolation.test.sh @@ -108,7 +108,7 @@ test_two_worktrees_same_leaf_get_distinct_compose_names() { # never trips teardown's dirty-worktree check or leaks into a crewmate commit. test_marker_is_git_excluded() { - local tmp proj fakebin wt out status_out + local tmp proj fakebin wt status_out tmp=$(fm_test_tmproot fm-spawn-compose-excl) proj=$(fm_git_init_commit "$tmp/proj" && printf '%s\n' "$tmp/proj") fakebin=$(make_spawn_fakebin "$tmp/fake") @@ -116,7 +116,7 @@ test_marker_is_git_excluded() { wt="$tmp/excltest-9ab1/3/excltest" git -C "$proj" worktree add -q --detach "$wt" || fail "setup: worktree add failed" - out=$(run_spawn "$tmp/home" excl-task-c "$proj" "$wt" "$fakebin") + run_spawn "$tmp/home" excl-task-c "$proj" "$wt" "$fakebin" >/dev/null expect_code 0 "$?" "spawn should succeed" assert_present "$wt/.treehouse-compose-project" "worktree missing the compose-project marker" From 77cfafae11f5250cea0eb47dcbe08609f65073ec Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 23:32:13 -0400 Subject: [PATCH 07/41] feat(spawn): auto-compact claude crewmates around 200k tokens Claude crewmates and secondmates are separate claude processes launched by fm-spawn.sh; they never inherit the primary's .claude/settings.local.json, so they only compact at the hard context wall instead of proactively. Add CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 to the claude launch template. The var is capped at the model's real context window, so a fixed 200000 lands auto-compaction around 200k regardless of whether the spawned model's window is 200k or (as verified live for --model sonnet on this account, which carries the 1M-context beta) ~1M. The primary's own two-var form (WINDOW=1000000 + PCT_OVERRIDE=20) assumes a 1M window and is not safe to copy blindly for a smaller-window model/account combination. grok, codex, opencode, and pi launches are left untouched: grok's binary is not verified to honor Claude Code env vars, and the other three ignore them harmlessly but should not carry a var meant for a harness they aren't. --- .agents/skills/harness-adapters/SKILL.md | 6 ++++++ bin/fm-spawn.sh | 22 ++++++++++++++++++++- tests/fm-backend-orca.test.sh | 2 +- tests/fm-spawn-dispatch-profile.test.sh | 25 +++++++++++++++++++++++- 4 files changed, 52 insertions(+), 3 deletions(-) diff --git a/.agents/skills/harness-adapters/SKILL.md b/.agents/skills/harness-adapters/SKILL.md index 63f378bb7..56210389c 100644 --- a/.agents/skills/harness-adapters/SKILL.md +++ b/.agents/skills/harness-adapters/SKILL.md @@ -132,6 +132,12 @@ Its broader dark-TRUECOLOR placeholder handling and dark-theme tradeoff are docu That styled capture is internal to the boolean detector only. `fm-peek` and every other human or LLM-facing capture path stays plain `tmux capture-pane` with no escape codes. +**Auto-compaction fact (verified 2026-07-09, Claude Code 2.1.206).** +`bin/fm-spawn.sh` launches every claude crewmate and secondmate with `CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000`, scoped the same way as `CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false` above, so these separately-launched processes auto-compact around 200k tokens the way the primary already does via its own `.claude/settings.local.json` (which they never inherit). +The var is read once at process startup and is capped at the model's real context window, so a fixed `200000` stays correct regardless of the spawned model's window: `min(200000, window)` is `200000` either way. +Verified live in a tmux pane: `claude --model sonnet` on this account resolves to a ~967k-token auto-compact window with no override (`/context` reported `Auto-compact window: 967k tokens`, i.e. the 1M-context beta, not the 200k a plain Sonnet window would imply); relaunched with `CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000` set, `/context` reported `39.5k/200k tokens (20%)` and `Auto-compact window: 200k tokens` - the window pinned to the override as expected. +This is why the primary's own two-var form (`CLAUDE_CODE_AUTO_COMPACT_WINDOW=1000000` + `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE=20`) is not copied here: that pair assumes a 1M window, and would silently cap a smaller-window model/account combination to a far more aggressive trigger. + **Primary-session guard fact (verified 2026-07-04, Claude Code 2.1.201; preserved 2026-07-08, Claude Code 2.1.204).** This is separate from the per-task crewmate turn-end hook above (that one just `touch`es a marker file in a task's own `.claude/settings.local.json`). The firstmate PRIMARY's own `.claude/settings.json` registers `bin/fm-turnend-guard.sh` as a Stop hook, and exiting with status 2 plus stderr reliably forces the model to continue. diff --git a/bin/fm-spawn.sh b/bin/fm-spawn.sh index 4a336c431..79a530a3b 100755 --- a/bin/fm-spawn.sh +++ b/bin/fm-spawn.sh @@ -85,6 +85,13 @@ # files below), and record it in meta as compose_project=, so two # worktrees of the same project never collide on Compose's default # project-naming (fm-brief.sh documents how a crewmate consumes it). +# claude launches (crewmate, scout, and secondmate) carry +# CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 so they auto-compact the way the primary +# interactive session already does via its own .claude/settings.local.json, which a +# separately-launched claude process never inherits; see the comment on the claude +# case in launch_template() below for why 200000 alone (not the primary's +# WINDOW=1000000 + PCT_OVERRIDE=20 pair) is the value that stays correct regardless +# of the spawned model's real context window. set -eu SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" @@ -327,7 +334,20 @@ launch_template() { # does NOT suppress the interactive ghost text (verified empirically), so the env # var is the correct control. The dim-aware composer reader in fm-tmux-lib.sh is # the defense-in-depth backstop for any pane this flag cannot reach. - claude) printf '%s' 'CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false claude --dangerously-skip-permissions __MODELFLAG____EFFORTFLAG__"$(cat __BRIEF__)"' ;; + # + # CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 gives claude crewmates and secondmates + # the same auto-compaction the primary interactive session already has via its + # own .claude/settings.local.json, which these separately-launched processes + # never inherit. The var is read once at startup and is CAPPED at the model's + # actual context window, so a single fixed value stays correct across whatever + # crewmate model is requested: min(200000, window) is 200000 whether the model's + # window is 200k or (as verified for --model sonnet on this account, which gets + # the 1M-context beta) ~1M, landing auto-compaction "around 200k" either way. + # Do not copy the primary's two-var form (WINDOW=1000000 + PCT_OVERRIDE=20): + # that pair assumes a 1M window and is not safe for a harness/account + # combination with a smaller one. See the harness-adapters skill's claude + # section for the dated /context verification. + claude) printf '%s' 'CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 claude --dangerously-skip-permissions __MODELFLAG____EFFORTFLAG__"$(cat __BRIEF__)"' ;; codex) if [ "$kind" = secondmate ]; then printf '%s' 'codex __MODELFLAG____EFFORTFLAG__--dangerously-bypass-approvals-and-sandbox "$(cat __BRIEF__)"' diff --git a/tests/fm-backend-orca.test.sh b/tests/fm-backend-orca.test.sh index 38fd4a9c5..d1029f1e3 100755 --- a/tests/fm-backend-orca.test.sh +++ b/tests/fm-backend-orca.test.sh @@ -502,7 +502,7 @@ test_spawn_writes_orca_metadata_and_launches_harness() { "spawn should reuse the implicit terminal returned by Orca worktree creation" assert_contains "$(cat "$log")" $'orca\x1f''terminal'$'\x1f''send'$'\x1f''--terminal'$'\x1f''term-spawn'$'\x1f''--text'$'\x1f''export GOTMPDIR=/tmp/fm-orcaspawnz1/gotmp'$'\x1f''--enter'$'\x1f''--json' \ "spawn did not export GOTMPDIR through the Orca terminal" - assert_contains "$(cat "$log")" "CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false claude --dangerously-skip-permissions" \ + assert_contains "$(cat "$log")" "CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 claude --dangerously-skip-permissions" \ "spawn did not send the selected harness launch command through Orca" rm -rf "/tmp/fm-$id" pass "fm-spawn.sh --backend orca: reuses implicit terminal, records metadata, launches harness" diff --git a/tests/fm-spawn-dispatch-profile.test.sh b/tests/fm-spawn-dispatch-profile.test.sh index d549c9bce..7e5b4b87c 100755 --- a/tests/fm-spawn-dispatch-profile.test.sh +++ b/tests/fm-spawn-dispatch-profile.test.sh @@ -118,7 +118,7 @@ test_no_profile_keeps_claude_launch_unchanged() { assert_meta_profile "$HOME_DIR/state/$id.meta" claude default default launch=$(cat "$LAUNCH_LOG") - expected="CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false claude --dangerously-skip-permissions \"\$(cat '$HOME_DIR/data/$id/brief.md')\"" + expected="CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 claude --dangerously-skip-permissions \"\$(cat '$HOME_DIR/data/$id/brief.md')\"" [ "$launch" = "$expected" ] || fail "no-profile claude launch changed"$'\n'"expected: $expected"$'\n'"actual: $launch" pass "no --model/--effort records defaults and keeps the claude launch byte-identical" } @@ -221,9 +221,31 @@ test_claude_threads_model_and_effort() { launch=$(cat "$LAUNCH_LOG") assert_contains "$launch" "claude --dangerously-skip-permissions --model 'sonnet' --effort 'high'" \ "claude launch did not thread model and effort flags" + assert_contains "$launch" "CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000" \ + "claude launch must carry the auto-compaction window env var regardless of model/effort flags" pass "claude receives --model and --effort profile flags" } +# Only claude launches carry the auto-compaction env var; codex, opencode, and pi +# ignore Claude Code env vars harmlessly, but firstmate should not set it for them +# (grok is intentionally out of scope too - unverified whether its binary honors it). +test_only_claude_launch_carries_autocompact_window() { + local rec id out status launch harness + for harness in codex opencode pi grok; do + id="profile-autocompact-$harness-z17" + rec=$(make_spawn_case "profile-autocompact-$harness" "$harness" "$id") + read_case_record "$rec" + + out=$(run_spawn "$HOME_DIR" "$WT_DIR" "$FAKEBIN_DIR" "$LAUNCH_LOG" "$id" "$PROJ_DIR") + status=$? + expect_code 0 "$status" "$harness spawn without profile flags should succeed" + launch=$(cat "$LAUNCH_LOG") + assert_not_contains "$launch" "CLAUDE_CODE_AUTO_COMPACT_WINDOW" \ + "$harness launch must not carry claude's auto-compaction env var" + done + pass "codex, opencode, pi, and grok launches never carry CLAUDE_CODE_AUTO_COMPACT_WINDOW" +} + test_codex_threads_model_and_effort() { local rec id out status launch id=profile-codex-z3 @@ -391,6 +413,7 @@ test_active_dispatch_profile_allows_explicit_harness test_active_dispatch_profile_allows_positional_harness test_active_dispatch_profile_allows_raw_launch_command test_claude_threads_model_and_effort +test_only_claude_launch_carries_autocompact_window test_codex_threads_model_and_effort test_codex_omits_invalid_max_effort test_grok_threads_model_and_reasoning_effort From f8fca83955f640a6e147ac1a90f585d18d25b79b Mon Sep 17 00:00:00 2001 From: Michael Date: Fri, 10 Jul 2026 00:12:35 -0400 Subject: [PATCH 08/41] no-mistakes(test): fix environment-fragile test failures (jq, git, harness detection) --- bin/backends/cmux.sh | 17 ++++++++++------- tests/fm-session-start.test.sh | 4 ++++ tests/fm-teardown.test.sh | 17 +++++++++++++++++ 3 files changed, 31 insertions(+), 7 deletions(-) diff --git a/bin/backends/cmux.sh b/bin/backends/cmux.sh index c37f003f4..478e9ee50 100644 --- a/bin/backends/cmux.sh +++ b/bin/backends/cmux.sh @@ -398,13 +398,16 @@ fm_backend_cmux_parse_target() { # # (fm_backend_zellij_pane_exists) rather than the design sketch's original # read-screen-based suggestion. fm_backend_cmux_surface_exists() { # - local wsid=$1 sfid=$2 raw - # Capture the CLI call's own exit status before piping to jq: on an empty - # stdin (the CLI call failed and printed nothing) `jq -e` still exits 0, so - # piping unconditionally would silently treat a failed call as "not found" - # rather than propagating the failure. - raw=$(fm_backend_cmux_cli list-panes --workspace "$wsid" --json --id-format uuids 2>/dev/null) || return 1 - printf '%s' "$raw" | jq -e --arg s "$sfid" '[.panes[]? | select(.surface_ids // [] | index($s))] | length > 0' >/dev/null 2>&1 + local wsid=$1 sfid=$2 out + # Capture the list-panes call and gate on its OWN exit/output before piping to + # jq: an absent target makes list-panes fail with no output, and `jq -e` on + # empty input exits 0 on jq 1.6 (only 1.7+ reports the no-result exit 4), so + # relying on the pipe's trailing jq status alone would false-positive the + # surface as present on older jq. A failed or empty list-panes means no surface. + out=$(fm_backend_cmux_cli list-panes --workspace "$wsid" --json --id-format uuids 2>/dev/null) || return 1 + [ -n "$out" ] || return 1 + printf '%s' "$out" \ + | jq -e --arg s "$sfid" '[.panes[]? | select(.surface_ids // [] | index($s))] | length > 0' >/dev/null 2>&1 } # fm_backend_cmux_target_ready: parse the target and verify it is live via diff --git a/tests/fm-session-start.test.sh b/tests/fm-session-start.test.sh index 86aeda132..a601d07e4 100755 --- a/tests/fm-session-start.test.sh +++ b/tests/fm-session-start.test.sh @@ -193,6 +193,10 @@ SH # (no ambient markers) still passes. run_session_start() { local home=$1 root=$2 path=$3 + # Unset the harness environment markers so the faked `ps` is the sole source of + # harness detection. Run from inside a real harness session (e.g. CLAUDECODE=1), + # fm-harness.sh's env-marker layer would otherwise short-circuit detection and + # override the harness the fake `ps` is meant to report. A no-op in a clean shell. env -u CLAUDECODE -u PI_CODING_AGENT -u GROK_AGENT \ FM_HOME="$home" FM_ROOT_OVERRIDE="$root" PATH="$path" \ "$SESSION_START" diff --git a/tests/fm-teardown.test.sh b/tests/fm-teardown.test.sh index 95b7c006c..8ceedd8cd 100755 --- a/tests/fm-teardown.test.sh +++ b/tests/fm-teardown.test.sh @@ -61,6 +61,15 @@ TMP_ROOT=$(fm_test_tmproot fm-teardown-tests) REAL_GIT_FOR_TEST=$(command -v git) export REAL_GIT_FOR_TEST +# fm-teardown.sh's content-in-default fallback merges via `git merge-tree +# --write-tree`, which git only grew in 2.38. On an older git the fallback is +# inconclusive by design (teardown refuses rather than guess), so the two content +# ALLOW cases below cannot pass. Skip them there instead of false-failing; a +# modern-git CI runs them fully. +git_has_merge_tree_write_tree() { + git merge-tree -h 2>&1 | grep -q -- '--write-tree' +} + # Build a fresh sandbox for one test case. Sets up: # $CASE/state/ - firstmate state dir (with a fresh watcher beacon) # $CASE/fakebin/ - mocks for treehouse, tmux (PATH-prepended by caller) @@ -804,6 +813,10 @@ test_pr_check_records_remote_head_when_local_lags() { test_content_in_default_fallback_allows() { local case_dir rc + if ! git_has_merge_tree_write_tree; then + pass "worktree whose content already landed in the default branch is torn down (content fallback) [skipped: git < 2.38 lacks merge-tree --write-tree]" + return 0 + fi case_dir=$(make_case content-landed) write_meta "$case_dir" no-mistakes ship # No pr= recorded and the default gh-axi mock reports no PR, so the merged-PR path @@ -824,6 +837,10 @@ test_content_in_default_fallback_allows() { test_content_fallback_refreshes_stale_origin_ref() { local case_dir rc + if ! git_has_merge_tree_write_tree; then + pass "content fallback refreshes origin default before comparing trees [skipped: git < 2.38 lacks merge-tree --write-tree]" + return 0 + fi case_dir=$(make_case content-stale-ref) write_meta "$case_dir" no-mistakes ship wt_commit_file "$case_dir" feature.txt hello "add feature" From 4d50cede597ced9d3ce2e05d83e6515d388f4142 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 13:23:52 -0400 Subject: [PATCH 09/41] feat(spawn): raise claude auto-compact window to 300000 Bump CLAUDE_CODE_AUTO_COMPACT_WINDOW from 200000 to 300000 for claude crewmates and secondmates, matching the primary session's own updated target. The override is capped at the model's real context window, so it is a no-op below 300000 and only takes effect above it; today's standard --model sonnet crewmate spawn resolves to the ~1M-context beta, well above the new threshold. Updated the fm-spawn.sh header comment, launch_template() comment, and the harness-adapters SKILL.md fact to describe this conditional behavior instead of the "regardless of window" framing that only held at the old 200000 value. --- .agents/skills/harness-adapters/SKILL.md | 9 ++++--- bin/fm-spawn.sh | 31 ++++++++++++++---------- tests/fm-backend-orca.test.sh | 2 +- tests/fm-spawn-dispatch-profile.test.sh | 4 +-- 4 files changed, 26 insertions(+), 20 deletions(-) diff --git a/.agents/skills/harness-adapters/SKILL.md b/.agents/skills/harness-adapters/SKILL.md index 56210389c..ec6bb5bc0 100644 --- a/.agents/skills/harness-adapters/SKILL.md +++ b/.agents/skills/harness-adapters/SKILL.md @@ -132,10 +132,11 @@ Its broader dark-TRUECOLOR placeholder handling and dark-theme tradeoff are docu That styled capture is internal to the boolean detector only. `fm-peek` and every other human or LLM-facing capture path stays plain `tmux capture-pane` with no escape codes. -**Auto-compaction fact (verified 2026-07-09, Claude Code 2.1.206).** -`bin/fm-spawn.sh` launches every claude crewmate and secondmate with `CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000`, scoped the same way as `CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false` above, so these separately-launched processes auto-compact around 200k tokens the way the primary already does via its own `.claude/settings.local.json` (which they never inherit). -The var is read once at process startup and is capped at the model's real context window, so a fixed `200000` stays correct regardless of the spawned model's window: `min(200000, window)` is `200000` either way. -Verified live in a tmux pane: `claude --model sonnet` on this account resolves to a ~967k-token auto-compact window with no override (`/context` reported `Auto-compact window: 967k tokens`, i.e. the 1M-context beta, not the 200k a plain Sonnet window would imply); relaunched with `CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000` set, `/context` reported `39.5k/200k tokens (20%)` and `Auto-compact window: 200k tokens` - the window pinned to the override as expected. +**Auto-compaction fact (verified 2026-07-09, Claude Code 2.1.206; shipped constant raised to 300000 2026-07-11).** +`bin/fm-spawn.sh` launches every claude crewmate and secondmate with `CLAUDE_CODE_AUTO_COMPACT_WINDOW=300000`, scoped the same way as `CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false` above, so these separately-launched processes auto-compact around 300k tokens the way the primary already does via its own `.claude/settings.local.json` (which they never inherit). +The var is read once at process startup and is capped at the model's real context window, so it can only lower the effective trigger, never raise it past the real ceiling: `min(300000, window)` is `300000` once `window` exceeds it, but a no-op (stays at `window`) below that. +This is not a universal guarantee for every possible future model choice - a differently-configured account/model capped at the plain ~200k window would see no change - but it is true for the window tier crew actually runs on today: `--model sonnet` on this account, the standard crewmate spawn, resolves to the ~1M-context beta, well above 300000. +Verified live in a tmux pane at the original 200000 value (the mechanism is unchanged, only the shipped constant moved to 300000): `claude --model sonnet` on this account resolves to a ~967k-token auto-compact window with no override (`/context` reported `Auto-compact window: 967k tokens`, i.e. the 1M-context beta, not the 200k a plain Sonnet window would imply); relaunched with `CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000` set, `/context` reported `39.5k/200k tokens (20%)` and `Auto-compact window: 200k tokens` - the window pinned to the override as expected. This is why the primary's own two-var form (`CLAUDE_CODE_AUTO_COMPACT_WINDOW=1000000` + `CLAUDE_AUTOCOMPACT_PCT_OVERRIDE=20`) is not copied here: that pair assumes a 1M window, and would silently cap a smaller-window model/account combination to a far more aggressive trigger. **Primary-session guard fact (verified 2026-07-04, Claude Code 2.1.201; preserved 2026-07-08, Claude Code 2.1.204).** diff --git a/bin/fm-spawn.sh b/bin/fm-spawn.sh index 79a530a3b..5f453c379 100755 --- a/bin/fm-spawn.sh +++ b/bin/fm-spawn.sh @@ -86,12 +86,12 @@ # worktrees of the same project never collide on Compose's default # project-naming (fm-brief.sh documents how a crewmate consumes it). # claude launches (crewmate, scout, and secondmate) carry -# CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 so they auto-compact the way the primary +# CLAUDE_CODE_AUTO_COMPACT_WINDOW=300000 so they auto-compact the way the primary # interactive session already does via its own .claude/settings.local.json, which a # separately-launched claude process never inherits; see the comment on the claude -# case in launch_template() below for why 200000 alone (not the primary's -# WINDOW=1000000 + PCT_OVERRIDE=20 pair) is the value that stays correct regardless -# of the spawned model's real context window. +# case in launch_template() below for why 300000 (not the primary's +# WINDOW=1000000 + PCT_OVERRIDE=20 pair) is safe to hardcode: it is capped at the +# real context window, so it is a no-op below that window and takes effect above it. set -eu SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" @@ -335,19 +335,24 @@ launch_template() { # var is the correct control. The dim-aware composer reader in fm-tmux-lib.sh is # the defense-in-depth backstop for any pane this flag cannot reach. # - # CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 gives claude crewmates and secondmates - # the same auto-compaction the primary interactive session already has via its - # own .claude/settings.local.json, which these separately-launched processes - # never inherit. The var is read once at startup and is CAPPED at the model's - # actual context window, so a single fixed value stays correct across whatever - # crewmate model is requested: min(200000, window) is 200000 whether the model's - # window is 200k or (as verified for --model sonnet on this account, which gets - # the 1M-context beta) ~1M, landing auto-compaction "around 200k" either way. + # CLAUDE_CODE_AUTO_COMPACT_WINDOW=300000 gives claude crewmates and secondmates + # a higher auto-compaction ceiling than the plain-window default, the same way + # the primary interactive session already has via its own + # .claude/settings.local.json, which these separately-launched processes never + # inherit. The var is read once at startup and is CAPPED at the model's actual + # context window, so it can only lower the effective trigger, never raise it past + # the real ceiling: min(300000, window) is 300000 once window exceeds it, but a + # no-op (stays at window) below that. This is not a universal guarantee for every + # possible future model choice - a differently-configured account/model capped at + # the plain ~200k window would see no change - but it is true for the window tier + # crew actually runs on today: as verified for --model sonnet on this account + # (the standard crewmate spawn), the window resolves to the ~1M-context beta, + # well above 300000. # Do not copy the primary's two-var form (WINDOW=1000000 + PCT_OVERRIDE=20): # that pair assumes a 1M window and is not safe for a harness/account # combination with a smaller one. See the harness-adapters skill's claude # section for the dated /context verification. - claude) printf '%s' 'CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 claude --dangerously-skip-permissions __MODELFLAG____EFFORTFLAG__"$(cat __BRIEF__)"' ;; + claude) printf '%s' 'CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=300000 claude --dangerously-skip-permissions __MODELFLAG____EFFORTFLAG__"$(cat __BRIEF__)"' ;; codex) if [ "$kind" = secondmate ]; then printf '%s' 'codex __MODELFLAG____EFFORTFLAG__--dangerously-bypass-approvals-and-sandbox "$(cat __BRIEF__)"' diff --git a/tests/fm-backend-orca.test.sh b/tests/fm-backend-orca.test.sh index d1029f1e3..d69128112 100755 --- a/tests/fm-backend-orca.test.sh +++ b/tests/fm-backend-orca.test.sh @@ -502,7 +502,7 @@ test_spawn_writes_orca_metadata_and_launches_harness() { "spawn should reuse the implicit terminal returned by Orca worktree creation" assert_contains "$(cat "$log")" $'orca\x1f''terminal'$'\x1f''send'$'\x1f''--terminal'$'\x1f''term-spawn'$'\x1f''--text'$'\x1f''export GOTMPDIR=/tmp/fm-orcaspawnz1/gotmp'$'\x1f''--enter'$'\x1f''--json' \ "spawn did not export GOTMPDIR through the Orca terminal" - assert_contains "$(cat "$log")" "CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 claude --dangerously-skip-permissions" \ + assert_contains "$(cat "$log")" "CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=300000 claude --dangerously-skip-permissions" \ "spawn did not send the selected harness launch command through Orca" rm -rf "/tmp/fm-$id" pass "fm-spawn.sh --backend orca: reuses implicit terminal, records metadata, launches harness" diff --git a/tests/fm-spawn-dispatch-profile.test.sh b/tests/fm-spawn-dispatch-profile.test.sh index 7e5b4b87c..68717c985 100755 --- a/tests/fm-spawn-dispatch-profile.test.sh +++ b/tests/fm-spawn-dispatch-profile.test.sh @@ -118,7 +118,7 @@ test_no_profile_keeps_claude_launch_unchanged() { assert_meta_profile "$HOME_DIR/state/$id.meta" claude default default launch=$(cat "$LAUNCH_LOG") - expected="CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000 claude --dangerously-skip-permissions \"\$(cat '$HOME_DIR/data/$id/brief.md')\"" + expected="CLAUDE_CODE_ENABLE_PROMPT_SUGGESTION=false CLAUDE_CODE_AUTO_COMPACT_WINDOW=300000 claude --dangerously-skip-permissions \"\$(cat '$HOME_DIR/data/$id/brief.md')\"" [ "$launch" = "$expected" ] || fail "no-profile claude launch changed"$'\n'"expected: $expected"$'\n'"actual: $launch" pass "no --model/--effort records defaults and keeps the claude launch byte-identical" } @@ -221,7 +221,7 @@ test_claude_threads_model_and_effort() { launch=$(cat "$LAUNCH_LOG") assert_contains "$launch" "claude --dangerously-skip-permissions --model 'sonnet' --effort 'high'" \ "claude launch did not thread model and effort flags" - assert_contains "$launch" "CLAUDE_CODE_AUTO_COMPACT_WINDOW=200000" \ + assert_contains "$launch" "CLAUDE_CODE_AUTO_COMPACT_WINDOW=300000" \ "claude launch must carry the auto-compaction window env var regardless of model/effort flags" pass "claude receives --model and --effort profile flags" } From 7f79f5d64421c0097719ae25c1fc716f46e91a0c Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 13:42:46 -0400 Subject: [PATCH 10/41] fix(teardown): recognize work fast-forwarded into local main as landed content_in_default() only ever diffed a branch against the fetched refs/remotes/origin/$name, so a no-mistakes-mode task whose branch was merged into the LOCAL default branch by hand (e.g. after its upstream PR was closed unmerged) was never recognized as landed and teardown refused it. It now also checks the local refs/heads/$name ref. --- bin/fm-teardown.sh | 55 ++++++++++++++++++++------------- tests/fm-teardown.test.sh | 65 +++++++++++++++++++++++++++++++++++++++ 2 files changed, 99 insertions(+), 21 deletions(-) diff --git a/bin/fm-teardown.sh b/bin/fm-teardown.sh index a7ec81570..f8929d5fe 100755 --- a/bin/fm-teardown.sh +++ b/bin/fm-teardown.sh @@ -274,27 +274,16 @@ pr_is_merged() { unpushed_patches_are_in_pr_head "$head" } -# Is the branch's content already present in the up-to-date default branch? Fetches -# first, then checks every path HEAD changed since its merge-base with the default -# branch: if the default branch already matches HEAD at each of those paths, HEAD -# introduces nothing the default branch does not already contain (e.g. its change -# landed via squash). This isolates branch-only changes, so unrelated commits the -# default branch gained past the merge-base do not count as "added" - equivalent to -# a 3-way merge's resulting tree matching the default branch's tree, but built from -# plain diff/merge-base so it works on git older than 2.38 (which is what `git -# merge-tree --write-tree` requires). Returns non-zero when inconclusive (no default -# ref, or a real difference), so the caller refuses rather than guesses. -content_in_default() { - local name ref base changed_file - name=$(default_branch) || return 1 - if git -C "$WT" remote get-url origin >/dev/null 2>&1; then - git -C "$WT" fetch --quiet origin "+refs/heads/$name:refs/remotes/origin/$name" >/dev/null 2>&1 || return 1 - ref="refs/remotes/origin/$name" - elif git -C "$WT" rev-parse --quiet --verify "refs/heads/$name" >/dev/null 2>&1; then - ref="refs/heads/$name" - else - return 1 - fi +# Does HEAD's content already match the given ref? Checks every path HEAD changed +# since its merge-base with ref: if ref already matches HEAD at each of those paths, +# HEAD introduces nothing ref does not already contain (e.g. its change landed via +# squash). This isolates branch-only changes, so unrelated commits ref gained past +# the merge-base do not count as "added" - equivalent to a 3-way merge's resulting +# tree matching ref's tree, but built from plain diff/merge-base so it works on git +# older than 2.38 (which is what `git merge-tree --write-tree` requires). Returns +# non-zero when inconclusive (ref missing, or a real difference). +content_matches_ref() { + local ref=$1 base changed_file git -C "$WT" rev-parse --quiet --verify "$ref^{commit}" >/dev/null 2>&1 || return 1 base=$(git -C "$WT" merge-base "$ref" HEAD 2>/dev/null) || return 1 local -a changed=() @@ -305,6 +294,30 @@ content_in_default() { git -C "$WT" diff --quiet "$ref" HEAD -- "${changed[@]}" 2>/dev/null } +# Is the branch's content already present in the up-to-date default branch? Checks +# two candidate refs, succeeding if either already contains HEAD's content: the +# fetched remote-tracking ref refs/remotes/origin/$name (the normal PR-merged case), +# and the LOCAL refs/heads/$name (a worktree and its project's primary checkout are +# linked worktrees of the same repository, so this local ref already reflects a +# firstmate-performed local fast-forward merge, even for a project whose delivery +# mode is not local-only - e.g. a PR closed unmerged upstream but the branch was +# landed into local main by hand on the captain's approval). A fetch failure only +# rules out the remote-tracking path; the local ref is still tried. Returns +# non-zero only when neither candidate ref exists or matches. +content_in_default() { + local name + name=$(default_branch) || return 1 + if git -C "$WT" remote get-url origin >/dev/null 2>&1; then + if git -C "$WT" fetch --quiet origin "+refs/heads/$name:refs/remotes/origin/$name" >/dev/null 2>&1; then + content_matches_ref "refs/remotes/origin/$name" && return 0 + fi + fi + if git -C "$WT" rev-parse --quiet --verify "refs/heads/$name" >/dev/null 2>&1; then + content_matches_ref "refs/heads/$name" && return 0 + fi + return 1 +} + # Has the worktree's committed work actually LANDED, though its commits are not # reachable from any remote-tracking branch? True when a merged PR proves the # current local work is contained in the PR head, OR the content is already in the diff --git a/tests/fm-teardown.test.sh b/tests/fm-teardown.test.sh index 8ceedd8cd..50a7aecea 100755 --- a/tests/fm-teardown.test.sh +++ b/tests/fm-teardown.test.sh @@ -38,6 +38,8 @@ # (o) fm-pr-check rerun after HEAD moved -> no stale pr_head # (p) fm-pr-check when local HEAD lags -> record remote PR head # (q) no-mistakes + NO pr= recorded, PR discovered by branch -> ALLOW (yolo/no-CI merge) +# (z) no-mistakes + PR closed unmerged, never pushed, merged +# into LOCAL default branch by hand -> ALLOW (local-land fix) # # Also covers backlog teardown-lock-race: a git index.lock left in the worktree by a # killed crew process (bin/fm-teardown.sh's teardown_treehouse_return). @@ -245,6 +247,37 @@ SH chmod +x "$case_dir/fakebin/gh-axi" "$case_dir/fakebin/gh" } +# Override GitHub lookups to report PR 7 as closed (unmerged) with the current +# worktree HEAD as its head - simulates a PR deliberately closed without merging +# (e.g. upstream repo not owned by this user). +add_gh_pr_closed_for_head() { + local case_dir=$1 head=$2 + cat > "$case_dir/fakebin/gh-axi" <<'SH' +#!/usr/bin/env bash +case "${1:-} ${2:-}" in + "pr list") + printf '%s\n' "count: 1 (showing first 1)" "pull_requests[1]{number,state}:" " 7,closed" ; exit 0 ;; + "pr view") + printf '%s\n' "pull_request:" " number: 7" " state: closed" ; exit 0 ;; +esac +exit 0 +SH + cat > "$case_dir/fakebin/gh" <&2 +exit 1 +SH + chmod +x "$case_dir/fakebin/gh-axi" "$case_dir/fakebin/gh" +} + append_pr_meta_for_current_head() { local case_dir=$1 head head=$(git -C "$case_dir/wt" rev-parse HEAD) @@ -636,6 +669,37 @@ test_no_mistakes_truly_unpushed_refuses() { pass "no-mistakes worktree with genuinely unlanded work is refused (safety preserved)" } +# Reproduces the local-land false-negative: a project registered under the normal +# no-mistakes delivery mode, whose task branch was never pushed to any remote and +# whose PR was deliberately closed unmerged (e.g. the upstream repo is not owned by +# this user), but which firstmate itself fast-forward-merged into the LOCAL default +# branch by hand, on explicit prior captain approval. content_in_default() used to +# only ever diff against the fetched refs/remotes/origin/$name, so it never saw this +# local-only landing outside of mode=local-only; it must now also accept containment +# in the LOCAL refs/heads/$name. +test_no_mistakes_merged_to_local_main_allows() { + local case_dir rc pr_head + case_dir=$(make_case nm-local-land) + write_meta "$case_dir" no-mistakes ship + wt_commit_file "$case_dir" feature.txt hello "shippable work" + pr_head=$(git -C "$case_dir/wt" rev-parse HEAD) + append_pr_meta_url "$case_dir" + add_gh_pr_closed_for_head "$case_dir" "$pr_head" + # Never pushed to origin or any fork. Instead, firstmate fast-forwards the + # project's LOCAL main to the worktree's HEAD, exactly as bin/fm-merge-local.sh + # would for an explicitly approved local landing. + git -C "$case_dir/project" update-ref refs/heads/main "$pr_head" + + set +e + run_teardown "$case_dir" > "$case_dir/stdout" 2> "$case_dir/stderr" + rc=$? + set -e + + expect_code 0 "$rc" "nm-local-land: teardown should succeed when work is merged into local main" + ! grep -q REFUSED "$case_dir/stderr" || fail "nm-local-land: teardown printed a REFUSED line" + pass "no-mistakes worktree with a closed PR but work merged into local main is torn down (local-land fix)" +} + test_squash_merged_branch_deleted_allows() { local case_dir rc pr_head case_dir=$(make_case squash-merged) @@ -1287,6 +1351,7 @@ test_local_only_truly_unpushed_refuses test_local_only_merged_to_local_main_allows test_no_mistakes_origin_remote_allows test_no_mistakes_truly_unpushed_refuses +test_no_mistakes_merged_to_local_main_allows test_local_only_force_overrides_unpushed test_herdr_teardown_clears_escalation_marker test_squash_merged_branch_deleted_allows From f75510eca52e6dc7e1ec596f55f15310ba2d9ac3 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 13:54:41 -0400 Subject: [PATCH 11/41] test(teardown): drop stale git-2.38 skip gate for content-in-default tests content_in_default()'s diff-based comparison (from local main's own prior rewrite) no longer depends on git merge-tree --write-tree, so the two tests that used to skip themselves on older git now run and pass here. --- tests/fm-teardown.test.sh | 17 ----------------- 1 file changed, 17 deletions(-) diff --git a/tests/fm-teardown.test.sh b/tests/fm-teardown.test.sh index 50a7aecea..c1cc2b35a 100755 --- a/tests/fm-teardown.test.sh +++ b/tests/fm-teardown.test.sh @@ -63,15 +63,6 @@ TMP_ROOT=$(fm_test_tmproot fm-teardown-tests) REAL_GIT_FOR_TEST=$(command -v git) export REAL_GIT_FOR_TEST -# fm-teardown.sh's content-in-default fallback merges via `git merge-tree -# --write-tree`, which git only grew in 2.38. On an older git the fallback is -# inconclusive by design (teardown refuses rather than guess), so the two content -# ALLOW cases below cannot pass. Skip them there instead of false-failing; a -# modern-git CI runs them fully. -git_has_merge_tree_write_tree() { - git merge-tree -h 2>&1 | grep -q -- '--write-tree' -} - # Build a fresh sandbox for one test case. Sets up: # $CASE/state/ - firstmate state dir (with a fresh watcher beacon) # $CASE/fakebin/ - mocks for treehouse, tmux (PATH-prepended by caller) @@ -877,10 +868,6 @@ test_pr_check_records_remote_head_when_local_lags() { test_content_in_default_fallback_allows() { local case_dir rc - if ! git_has_merge_tree_write_tree; then - pass "worktree whose content already landed in the default branch is torn down (content fallback) [skipped: git < 2.38 lacks merge-tree --write-tree]" - return 0 - fi case_dir=$(make_case content-landed) write_meta "$case_dir" no-mistakes ship # No pr= recorded and the default gh-axi mock reports no PR, so the merged-PR path @@ -901,10 +888,6 @@ test_content_in_default_fallback_allows() { test_content_fallback_refreshes_stale_origin_ref() { local case_dir rc - if ! git_has_merge_tree_write_tree; then - pass "content fallback refreshes origin default before comparing trees [skipped: git < 2.38 lacks merge-tree --write-tree]" - return 0 - fi case_dir=$(make_case content-stale-ref) write_meta "$case_dir" no-mistakes ship wt_commit_file "$case_dir" feature.txt hello "add feature" From e77f866c3f8d8b2cbfe876c4ee5c00bb8edc53e6 Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 17:10:29 -0400 Subject: [PATCH 12/41] feat(dispatch): add model/effort tiers and mechanical quality guardrails Extends config/crew-dispatch.json's schema with model/effort/ultracode axes per the resource-tiering design: docs/examples/crew-dispatch.json gains rules for the firstmate-repo supervision-backbone carve-out, safety-critical/security/migration work (opus/xhigh, ultracode-flagged), and architectural/product tradeoffs (opus/xhigh). Adds the four hardened guardrails as mechanical checks rather than prose aspirations: - bin/fm-tier-guard.sh: firstmate-observable escalation triggers (diff size and elapsed time vs. the trivial tier's envelope, plus a general heavy-scale ceiling for any tier). - bin/fm-risk-tripwire.sh: a keyword/path trip-wire against a task's brief and diff, independent of the natural-language dispatch match. - bin/fm-ultracode-guard.sh: flag/reviewed/check subcommands that mechanically confirm a genuinely independent second pass ran before an ultracode-flagged task reaches PR-ready. - A documented model-verification gate in AGENTS.md section 4, extending the existing unverified-harness discipline to model names. Verification tiering (light/standard/heavy no-mistakes runs) is deliberately out of scope here and left for a follow-up change. --- AGENTS.md | 15 +++ bin/fm-risk-tripwire.sh | 121 +++++++++++++++++++ bin/fm-tier-guard.sh | 121 +++++++++++++++++++ bin/fm-ultracode-guard.sh | 100 ++++++++++++++++ docs/configuration.md | 3 +- docs/examples/crew-dispatch.json | 15 +++ tests/fm-risk-tripwire.test.sh | 146 +++++++++++++++++++++++ tests/fm-tier-guard.test.sh | 192 +++++++++++++++++++++++++++++++ tests/fm-ultracode-guard.test.sh | 168 +++++++++++++++++++++++++++ 9 files changed, 880 insertions(+), 1 deletion(-) create mode 100755 bin/fm-risk-tripwire.sh create mode 100755 bin/fm-tier-guard.sh create mode 100755 bin/fm-ultracode-guard.sh create mode 100644 tests/fm-risk-tripwire.test.sh create mode 100644 tests/fm-tier-guard.test.sh create mode 100644 tests/fm-ultracode-guard.test.sh diff --git a/AGENTS.md b/AGENTS.md index d51f22ee7..535398658 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -84,6 +84,12 @@ When dispatch profiles exist, consult them at every crewmate or scout intake and Routing precedence is an explicit per-task captain override, then the best-fit configured rule, then the configured default, then the static crewmate harness. The generic effort fallback and its precedence are owned by `harness-adapters`: explicit captain and standing configured effort win; otherwise use low for well-understood explicit work, xhigh for ambiguous investigation or design, intermediate levels proportionally, and never max without explicit captain preference. Do not add model-specific versions of that policy. +When a resolved profile sets `ultracode` (schema in `docs/configuration.md` "Crew dispatch profiles"), run `bin/fm-ultracode-guard.sh flag ` right after spawn so section 7's Validate gate is tracked mechanically rather than resting on memory. + +**Risk floor.** Run `bin/fm-risk-tripwire.sh ` before spawn and again once the task has a brief or a diff. +A hit floors the model/effort to the safety-critical profile (`opus`/`xhigh`, ultracode `independent-review`) regardless of which rule matched, and a hit found only at Validate time still floors the task in place and is never silently downgraded afterward. + +Before a dispatch rule may name a specific non-default model, verify it end-to-end on a trivial supervised task, exactly as this section already requires before naming a new harness adapter; do not use an unverified model live. `secondmate-provisioning` owns secondmate harness pins and config inheritance, while `harness-adapters` owns the harness consequences. Dispatch only on a backend that `fm-spawn` validates as spawn-capable. @@ -206,6 +212,15 @@ Running, fixing, or CI states remain working; parked approval or fix-review stat A worker hand-editing, committing, aborting, or restarting during an active validation run duplicates pipeline ownership; steer it back to the gate response flow. The worker reports the PR when CI first becomes green rather than waiting for merge monitoring to finish. +**Escalation triggers (mechanical, never rest on self-report alone).** +For a task assigned the trivial tier (Haiku/low), run `bin/fm-tier-guard.sh ` during Validate, or whenever a heartbeat review touches it, to check whether its diff or elapsed time has outgrown that tier's envelope; any tier also escalates once its diff crosses the script's general heavy-scale ceiling. +A report escalates the task to at least Sonnet/high in place, without losing its branch or context, mirroring `bin/fm-promote.sh`'s in-place promotion. +A crewmate's own report - it cannot find root cause, a "confirmed" fix touches a shared path, or it raised a tradeoff buried in seemingly mechanical work - is a second, non-mechanical trigger with the same effect. +Either way, escalate the model/effort in place and never silently de-escalate for the rest of that task's life. + +**Ultracode confirmation.** +Before advancing an ultracode-flagged task to PR-ready, run `bin/fm-ultracode-guard.sh check `; it refuses until a genuinely separate task - dispatched independently, never a sub-task the flagged crewmate spawned itself - has reviewed the finished diff and its findings were addressed, recorded with `bin/fm-ultracode-guard.sh reviewed `. + ### PR ready, landing, and teardown For a ready PR, use `bin/fm-pr-check.sh` to record the PR and authoritative head and to arm merge monitoring. diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh new file mode 100755 index 000000000..8883617f9 --- /dev/null +++ b/bin/fm-risk-tripwire.sh @@ -0,0 +1,121 @@ +#!/usr/bin/env bash +# Mechanical risk trip-wire (guardrail #2): greps a task's brief text and, once +# code exists, its changed file paths for migration/auth/schema/security +# signals - a second, structurally different check from the natural-language +# "when" match a crew-dispatch rule already made, so a misclassified risky task +# cannot slip to a cheap model/effort tier on a single judgment call +# (data/research-resource-tiering-synthesis.md). +# +# Usage: fm-risk-tripwire.sh +# +# Checks whatever is available for : +# - data//brief.md, if it exists (works before spawn, right after +# bin/fm-brief.sh scaffolds it - the first checkpoint) +# - the worktree's changed file paths vs its project's default branch, if +# state/.meta records worktree=/project= (works after spawn, at +# Validate time - the second, binding checkpoint against the real diff) +# Errors (exit 2) if neither is available; that means there is nothing to +# check yet. +# +# Prints one "RISK: " line per hit and exits 1 if any fired; exits 0 +# (silent) when neither surface matched. A hit means: floor this task's +# model/effort to the safety-critical profile (opus/xhigh, ultracode +# independent-review) regardless of which rule the natural-language dispatch +# match picked, per AGENTS.md section 4's risk floor. +# +# This is a coarse, unpushed-diff-vs-default-branch name-only comparison, not +# the PR-aware exact diff bin/fm-review-diff.sh computes - good enough for a +# keyword scan, not a substitute for that script's authoritative base. +set -eu + +SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" +FM_ROOT="${FM_ROOT_OVERRIDE:-$(cd "$SCRIPT_DIR/.." && pwd)}" +FM_HOME="${FM_HOME:-${FM_ROOT_OVERRIDE:-$FM_ROOT}}" +STATE="${FM_STATE_OVERRIDE:-$FM_HOME/state}" +DATA="${FM_DATA_OVERRIDE:-$FM_HOME/data}" + +# shellcheck source=bin/fm-tangle-lib.sh +. "$SCRIPT_DIR/fm-tangle-lib.sh" + +usage() { + echo "usage: fm-risk-tripwire.sh " >&2 +} + +if [ "${1:-}" = "-h" ] || [ "${1:-}" = "--help" ]; then + usage + exit 0 +fi + +ID=${1:-} +[ -n "$ID" ] || { usage; exit 1; } +[ $# -le 1 ] || { usage; exit 1; } + +# Case-insensitive substring match against the brief's prose. Permissive by +# design (a false positive just means "double-check"; a false negative is the +# dangerous direction for a safety floor). +TEXT_KEYWORDS='auth|authentication|authorization|session|credential|password|secret|token|payment|billing|migration|schema|security|encrypt|decrypt|permission|access.control|data.deletion|bulk.mutation|public.exposure|breaking.change' + +# Substring match against one changed file's path (already lowercased by the caller). +path_is_risky() { + case "$1" in + .github/workflows/*|*/.github/workflows/*) return 0 ;; + bin/*|*/bin/*) return 0 ;; + *dockerfile*|*docker-compose*) return 0 ;; + *auth*|*migrat*|*schema*|*secret*|*credential*|*payment*|*billing*|*security*|*session*) return 0 ;; + esac + return 1 +} + +FOUND=0 +CHECKED=0 + +BRIEF="$DATA/$ID/brief.md" +if [ -f "$BRIEF" ]; then + CHECKED=1 + hit=$(grep -Eoi "$TEXT_KEYWORDS" "$BRIEF" | tr '[:upper:]' '[:lower:]' | sort -u | tr '\n' ',' | sed 's/,$//') + if [ -n "$hit" ]; then + echo "RISK: brief for $ID mentions risk-adjacent term(s): $hit" + FOUND=1 + fi +fi + +META="$STATE/$ID.meta" +if [ -f "$META" ]; then + WT=$(grep '^worktree=' "$META" | tail -1 | cut -d= -f2- || true) + PROJ=$(grep '^project=' "$META" | tail -1 | cut -d= -f2- || true) + if [ -n "$WT" ] && [ -n "$PROJ" ] && [ -d "$WT" ] && [ -d "$PROJ" ]; then + CHECKED=1 + DEFAULT=$(fm_default_branch "$PROJ" 2>/dev/null || true) + if [ -n "$DEFAULT" ]; then + BASE="$DEFAULT" + if git -C "$PROJ" remote get-url origin >/dev/null 2>&1; then + git -C "$WT" fetch origin "+refs/heads/$DEFAULT:refs/remotes/origin/$DEFAULT" --quiet 2>/dev/null || true + git -C "$WT" rev-parse --verify --quiet "refs/remotes/origin/$DEFAULT^{commit}" >/dev/null 2>&1 && BASE="origin/$DEFAULT" + fi + if git -C "$WT" rev-parse --verify --quiet "$BASE^{commit}" >/dev/null 2>&1; then + diff_paths=$(git -C "$WT" diff --name-only "$BASE...HEAD" -- 2>/dev/null || true) + risky_paths= + if [ -n "$diff_paths" ]; then + while IFS= read -r path; do + [ -n "$path" ] || continue + lower_path=$(printf '%s' "$path" | tr '[:upper:]' '[:lower:]') + if path_is_risky "$lower_path"; then + risky_paths="${risky_paths}${risky_paths:+, }$path" + fi + done <<< "$diff_paths" + fi + if [ -n "$risky_paths" ]; then + echo "RISK: diff for $ID touches risk-adjacent path(s): $risky_paths" + FOUND=1 + fi + fi + fi + fi +fi + +if [ "$CHECKED" -eq 0 ]; then + echo "error: neither $BRIEF nor a usable worktree/project in $META was found for $ID; nothing to check" >&2 + exit 2 +fi + +exit "$FOUND" diff --git a/bin/fm-tier-guard.sh b/bin/fm-tier-guard.sh new file mode 100755 index 000000000..be7effca2 --- /dev/null +++ b/bin/fm-tier-guard.sh @@ -0,0 +1,121 @@ +#!/usr/bin/env bash +# Escalation-trigger guard for the trivial (Haiku/low) dispatch tier: mechanically +# checks whether a task's actual diff size or elapsed time has outgrown the +# resource envelope its assigned model/effort tier implies, so under-resourcing +# is never caught by the crewmate's own self-report alone (guardrail #1, +# data/research-resource-tiering-synthesis.md). Read-only: never edits meta, +# the worktree, or the branch. +# +# Usage: fm-tier-guard.sh +# +# Reads state/.meta for model=/effort=, and that file's own mtime as a +# spawn-time proxy. Reuses bin/fm-review-diff.sh --stat for the actual diff size +# instead of re-deriving the authoritative base/PR-head resolution here. +# +# Prints one "ESCALATE: " line per triggered condition and exits 1 if +# any fired; exits 0 (silent) when the task is within its tier's envelope. +# Only the trivial (Haiku/low) tier has a size/age ceiling checked here; every +# other tier is covered only by the general, tier-independent ceiling below. +# On an escalation, bump the task's model/effort in place per AGENTS.md +# section 7's Validate step; never silently de-escalate afterward. +# +# Env overrides (starting points from +# data/scout-nomistakes-tiering-t8/report.md section 3, not yet calibrated +# against this fleet's own diff-size history): +# FM_TIER_TRIVIAL_MAX_LINES default 30 +# FM_TIER_TRIVIAL_MAX_FILES default 2 +# FM_TIER_TRIVIAL_MAX_AGE_SECONDS default 1800 (30 minutes) +# FM_TIER_HEAVY_MIN_LINES default 400 +# FM_TIER_HEAVY_MIN_FILES default 8 +set -eu + +SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" +FM_ROOT="${FM_ROOT_OVERRIDE:-$(cd "$SCRIPT_DIR/.." && pwd)}" +FM_HOME="${FM_HOME:-${FM_ROOT_OVERRIDE:-$FM_ROOT}}" +STATE="${FM_STATE_OVERRIDE:-$FM_HOME/state}" + +TRIVIAL_MAX_LINES=${FM_TIER_TRIVIAL_MAX_LINES:-30} +TRIVIAL_MAX_FILES=${FM_TIER_TRIVIAL_MAX_FILES:-2} +TRIVIAL_MAX_AGE=${FM_TIER_TRIVIAL_MAX_AGE_SECONDS:-1800} +HEAVY_MIN_LINES=${FM_TIER_HEAVY_MIN_LINES:-400} +HEAVY_MIN_FILES=${FM_TIER_HEAVY_MIN_FILES:-8} + +usage() { + echo "usage: fm-tier-guard.sh " >&2 +} + +if [ "${1:-}" = "-h" ] || [ "${1:-}" = "--help" ]; then + usage + exit 0 +fi + +ID=${1:-} +[ -n "$ID" ] || { usage; exit 1; } +[ $# -le 1 ] || { usage; exit 1; } + +META="$STATE/$ID.meta" +[ -f "$META" ] || { echo "error: no meta for task $ID at $META" >&2; exit 1; } + +MODEL=$(grep '^model=' "$META" | tail -1 | cut -d= -f2- || true) +EFFORT=$(grep '^effort=' "$META" | tail -1 | cut -d= -f2- || true) + +# Portable mtime; Linux stat lacks -f, macOS stat lacks -c (mirrors +# bin/fm-supervision-lib.sh's fm_sup_stat_mtime; not sourced from there to keep +# this script's only dependency on that library's actual contract-bearing +# function, fm_supervision_status, out of scope here). +tier_guard_stat_mtime() { + if [ "$(uname)" = Darwin ]; then + stat -f %m "$1" 2>/dev/null + else + stat -c %Y "$1" 2>/dev/null + fi +} + +is_trivial_tier() { + case "$MODEL" in + *haiku*) [ "$EFFORT" = low ] && return 0 ;; + esac + return 1 +} + +STAT_OUT=$("$SCRIPT_DIR/fm-review-diff.sh" "$ID" --stat 2>/dev/null) || { + echo "error: fm-review-diff.sh failed for $ID" >&2 + exit 1 +} + +FILES=0 +LINES=0 +if ! printf '%s\n' "$STAT_OUT" | grep -q '^no changes vs'; then + FILES=$(printf '%s\n' "$STAT_OUT" | grep -oE '[0-9]+ files? changed' | grep -oE '^[0-9]+' || true) + INS=$(printf '%s\n' "$STAT_OUT" | grep -oE '[0-9]+ insertions?\(\+\)' | grep -oE '^[0-9]+' || true) + DEL=$(printf '%s\n' "$STAT_OUT" | grep -oE '[0-9]+ deletions?\(-\)' | grep -oE '^[0-9]+' || true) + FILES=${FILES:-0} + INS=${INS:-0} + DEL=${DEL:-0} + LINES=$((INS + DEL)) +fi + +FOUND=0 + +if is_trivial_tier; then + if [ "$FILES" -gt "$TRIVIAL_MAX_FILES" ] || [ "$LINES" -gt "$TRIVIAL_MAX_LINES" ]; then + echo "ESCALATE: trivial (haiku/low) task $ID has grown to $FILES files / $LINES changed lines, past the trivial envelope ($TRIVIAL_MAX_FILES files / $TRIVIAL_MAX_LINES lines) - bump to at least sonnet/high" + FOUND=1 + fi + AGE_SRC=$(tier_guard_stat_mtime "$META" || true) + if [ -n "$AGE_SRC" ]; then + NOW=$(date +%s) + AGE=$((NOW - AGE_SRC)) + if [ "$AGE" -gt "$TRIVIAL_MAX_AGE" ]; then + echo "ESCALATE: trivial (haiku/low) task $ID has run ${AGE}s, past the trivial tier's ${TRIVIAL_MAX_AGE}s ceiling - bump to at least sonnet/high" + FOUND=1 + fi + fi +fi + +if [ "$FILES" -gt "$HEAVY_MIN_FILES" ] || [ "$LINES" -gt "$HEAVY_MIN_LINES" ]; then + echo "ESCALATE: task $ID's diff ($FILES files / $LINES changed lines) exceeds the heavy-scale ceiling ($HEAVY_MIN_FILES files / $HEAVY_MIN_LINES lines) regardless of its assigned tier - re-check whether it still matches its dispatched rule" + FOUND=1 +fi + +exit "$FOUND" diff --git a/bin/fm-ultracode-guard.sh b/bin/fm-ultracode-guard.sh new file mode 100755 index 000000000..099ae8edc --- /dev/null +++ b/bin/fm-ultracode-guard.sh @@ -0,0 +1,100 @@ +#!/usr/bin/env bash +# Ultracode enforcement (guardrail #3): mechanically confirms a genuinely +# independent second pass ran on an ultracode-flagged task's finished diff +# before it can go PR-ready - not a sub-task the same crewmate spawned itself +# (data/research-resource-tiering-synthesis.md). Tracks state via a plain +# marker file, state/.ultracode, the same convention as other +# firstmate state/ markers (state/.afk, state/.turn-ended); it never +# touches fm-spawn.sh or the task's own meta. +# +# Usage: +# fm-ultracode-guard.sh flag [] +# Records that was dispatched under a crew-dispatch rule whose +# resolved profile set ultracode=true. defaults to +# "independent-review" (the only role this fleet's rules use today; see +# docs/examples/crew-dispatch.json). Firstmate runs this right after +# spawning the task. +# fm-ultracode-guard.sh reviewed +# Records that - a distinct, separately dispatched task +# (its own state/.meta must exist) - independently +# reviewed 's finished diff and its findings were addressed. +# Refuses if equals or has no recorded meta, +# so a sub-task the same crewmate spawned itself cannot satisfy this. +# fm-ultracode-guard.sh check +# Exits 0 if was never flagged, or was flagged and reviewed. +# Exits 1 with an explanatory message if flagged but not yet reviewed - +# firstmate runs this before treating an ultracode-flagged task as +# PR-ready (AGENTS.md section 7's Validate step). +set -eu + +SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" +FM_ROOT="${FM_ROOT_OVERRIDE:-$(cd "$SCRIPT_DIR/.." && pwd)}" +FM_HOME="${FM_HOME:-${FM_ROOT_OVERRIDE:-$FM_ROOT}}" +STATE="${FM_STATE_OVERRIDE:-$FM_HOME/state}" + +usage() { + echo "usage: fm-ultracode-guard.sh flag []" >&2 + echo " fm-ultracode-guard.sh reviewed " >&2 + echo " fm-ultracode-guard.sh check " >&2 +} + +if [ "${1:-}" = "-h" ] || [ "${1:-}" = "--help" ]; then + usage + exit 0 +fi + +CMD=${1:-} +ID=${2:-} +[ -n "$CMD" ] && [ -n "$ID" ] || { usage; exit 1; } + +MARKER="$STATE/$ID.ultracode" + +cmd_flag() { + local role=${3:-independent-review} + [ $# -le 3 ] || { usage; exit 1; } + mkdir -p "$STATE" + # Overwrites any existing marker: re-flagging (e.g. after an escalation) + # deliberately clears a prior reviewed_by=, since that review was against an + # earlier version of the diff and the requirement starts over. + { + echo "role=$role" + } > "$MARKER" + echo "flagged $ID ultracode role=$role" +} + +cmd_reviewed() { + local reviewer=${3:-} + [ $# -eq 3 ] || { usage; exit 1; } + [ -n "$reviewer" ] || { echo "error: reviewed requires a reviewer-task-id" >&2; exit 1; } + [ -f "$MARKER" ] || { echo "error: $ID is not ultracode-flagged (no $MARKER); nothing to mark reviewed" >&2; exit 1; } + if [ "$reviewer" = "$ID" ]; then + echo "error: reviewer-task-id must be a task distinct from $ID - a task cannot independently review itself" >&2 + exit 1 + fi + if [ ! -f "$STATE/$reviewer.meta" ]; then + echo "error: $reviewer has no recorded state/$reviewer.meta - it must be a genuinely, separately dispatched task, not a made-up id or a sub-task $ID spawned itself" >&2 + exit 1 + fi + grep -qx "reviewed_by=$reviewer" "$MARKER" 2>/dev/null || echo "reviewed_by=$reviewer" >> "$MARKER" + echo "recorded $reviewer as the independent review of $ID" +} + +cmd_check() { + [ $# -eq 2 ] || { usage; exit 1; } + if [ ! -f "$MARKER" ]; then + exit 0 + fi + if grep -q '^reviewed_by=' "$MARKER" 2>/dev/null; then + exit 0 + fi + role=$(grep '^role=' "$MARKER" | tail -1 | cut -d= -f2- || true) + echo "error: $ID is ultracode-flagged (role=${role:-independent-review}) but has no recorded independent review yet - dispatch a genuinely separate task to review the finished diff, then run: fm-ultracode-guard.sh reviewed $ID " >&2 + exit 1 +} + +case "$CMD" in + flag) cmd_flag "$@" ;; + reviewed) cmd_reviewed "$@" ;; + check) cmd_check "$@" ;; + *) usage; exit 1 ;; +esac diff --git a/docs/configuration.md b/docs/configuration.md index 0d45d09d6..c9f459e8f 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -201,7 +201,7 @@ This section is the single owner of the canonical schema and its per-field seman { "when": "", "use": [ - { "harness": "", "model": "", "effort": "" } + { "harness": "", "model": "", "effort": "", "ultracode": "", "ultracode_role": "" } ], "select": "", "why": "" @@ -214,6 +214,7 @@ This section is the single owner of the canonical schema and its per-field seman Per rule, `when` and `use` are required. `use` may be a single profile object or an ordered array of profile objects; the single-object form stays fully backward-compatible, and every profile needs `harness`. `use.model`, `use.effort`, and `why` are optional. +`use.ultracode` and `use.ultracode_role` are optional: a profile with `ultracode: true` means a task matched to this rule must get a genuinely independent second pass on its finished diff before PR-ready, launched as its own separately dispatched task rather than a sub-task the implementing crewmate spawns itself, and `ultracode_role` is a short label for what that pass should do (`independent-review` is the only role this fleet's own rules use today); `bin/fm-ultracode-guard.sh` tracks this mechanically, and `AGENTS.md` section 4 and section 7's Validate step cover when it is flagged and gated. `select` is optional and currently supports `quota-balanced`. Absent `select` means use the first array element, or the only object in the single-object form; the first array element is the deterministic tie-break and the ultimate fallback. `default` is optional. diff --git a/docs/examples/crew-dispatch.json b/docs/examples/crew-dispatch.json index 8aec19650..8f27118e6 100644 --- a/docs/examples/crew-dispatch.json +++ b/docs/examples/crew-dispatch.json @@ -18,6 +18,21 @@ ], "select": "quota-balanced", "why": "Use a strong coding profile for broad design and implementation work." + }, + { + "when": "The task changes this firstmate repo's own supervision backbone: bin/fm-watch.sh, bin/fm-guard.sh, a harness turnend-guard hook, or other logic the always-on watcher/guard rely on for correctness.", + "use": { "harness": "claude", "model": "sonnet", "effort": "xhigh" }, + "why": "A narrow set of scripts backs every other task's supervision; a subtle regression here is silent and fleet-wide, so it earns deeper reasoning than the general firstmate-repo default even though the model tier does not need to change." + }, + { + "when": "Safety-critical, security-sensitive, or hard-to-reverse work: schema or data migrations, auth/session/credential changes, payment or billing logic, or anything touching a frozen or safety-critical spec.", + "use": { "harness": "claude", "model": "opus", "effort": "xhigh", "ultracode": true, "ultracode_role": "independent-review" }, + "why": "A model reviewing its own irreversible change has anchoring bias; an independent pass and the strongest available model are both cheap relative to the cost of a bad migration or a broken safety-flow." + }, + { + "when": "A genuine architectural or product tradeoff with long-term consequences that the crewmate must reason through and justify, not just implement.", + "use": { "harness": "claude", "model": "opus", "effort": "xhigh" }, + "why": "The deliverable's entire value is the quality of the argued judgment call; this is low-volume, high-stakes work, so the cost delta of the top model is bounded and well spent." } ], "default": { "harness": "codex", "model": "gpt-5.5", "effort": "medium" } diff --git a/tests/fm-risk-tripwire.test.sh b/tests/fm-risk-tripwire.test.sh new file mode 100644 index 000000000..fc4999fdc --- /dev/null +++ b/tests/fm-risk-tripwire.test.sh @@ -0,0 +1,146 @@ +#!/usr/bin/env bash +# Tests for bin/fm-risk-tripwire.sh: a brief mentioning a risk-adjacent term, +# or a diff touching a risk-adjacent path, must trip the wire regardless of +# how the task's dispatch rule classified it - the mechanical, structurally +# different check behind AGENTS.md section 4's risk floor. +set -u + +# shellcheck source=tests/lib.sh +. "$(dirname "${BASH_SOURCE[0]}")/lib.sh" +fm_git_identity fmtest fmtest@example.invalid + +TRIPWIRE="$ROOT/bin/fm-risk-tripwire.sh" +TMP_ROOT=$(fm_test_tmproot fm-risk-tripwire-tests) + +make_case() { + local name=$1 case_dir + case_dir="$TMP_ROOT/$name" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + + git init -q --bare "$case_dir/origin.git" + git -C "$case_dir/origin.git" symbolic-ref HEAD refs/heads/main + git clone -q "$case_dir/origin.git" "$case_dir/_seed" 2>/dev/null + printf 'base\n' > "$case_dir/_seed/feature.txt" + git -C "$case_dir/_seed" add feature.txt + git -C "$case_dir/_seed" -c user.email=t@t -c user.name=t commit -qm "origin baseline" + git -C "$case_dir/_seed" push -q origin main + rm -rf "$case_dir/_seed" + + git clone -q "$case_dir/origin.git" "$case_dir/project" + git -C "$case_dir/project" remote set-head origin main 2>/dev/null || true + git -C "$case_dir/project" worktree add -q -b "fm/$name" "$case_dir/wt" main + + printf '%s\n' "$case_dir" +} + +write_task_meta() { + local case_dir=$1 + fm_write_meta "$case_dir/state/task-x1.meta" \ + "window=fm-task-x1" \ + "worktree=$case_dir/wt" \ + "project=$case_dir/project" +} + +run_tripwire() { + local case_dir=$1 + FM_ROOT_OVERRIDE="$ROOT" \ + FM_STATE_OVERRIDE="$case_dir/state" \ + FM_DATA_OVERRIDE="$case_dir/data" \ + "$TRIPWIRE" task-x1 +} + +test_clean_brief_and_diff_passes() { + local case_dir out status + case_dir=$(make_case clean) + printf 'Add a --json flag to the status command.\n' > "$case_dir/data/task-x1/brief.md" + printf 'ordinary change\n' > "$case_dir/wt/feature.txt" + git -C "$case_dir/wt" add feature.txt + git -C "$case_dir/wt" commit -qm "ordinary change" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 0 "$status" "clean: an ordinary brief and diff must not trip the wire" + [ -z "$out" ] || fail "clean: expected no RISK output, got: $out" + pass "fm-risk-tripwire passes a clean brief and diff" +} + +test_brief_keyword_trips_wire() { + local case_dir out status + case_dir=$(make_case brief-keyword) + printf 'Add a data migration for the new billing schema.\n' > "$case_dir/data/task-x1/brief.md" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 1 "$status" "brief-keyword: a risk-worded brief must trip the wire" + assert_contains "$out" "RISK: brief for task-x1 mentions risk-adjacent term(s)" "brief-keyword: should name the brief hit" + assert_contains "$out" "migration" "brief-keyword: should surface the matched term" + pass "fm-risk-tripwire trips on a brief that mentions risk-adjacent terms" +} + +test_diff_path_trips_wire() { + local case_dir out status + case_dir=$(make_case diff-path) + printf 'Fix a typo in the help text.\n' > "$case_dir/data/task-x1/brief.md" + mkdir -p "$case_dir/wt/lib/auth" + printf 'session handling\n' > "$case_dir/wt/lib/auth/session.rb" + git -C "$case_dir/wt" add lib/auth/session.rb + git -C "$case_dir/wt" commit -qm "touch auth session code" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 1 "$status" "diff-path: a diff touching an auth path must trip the wire" + assert_contains "$out" "RISK: diff for task-x1 touches risk-adjacent path(s)" "diff-path: should name the diff hit" + assert_contains "$out" "lib/auth/session.rb" "diff-path: should list the risky path" + pass "fm-risk-tripwire trips on a diff that touches an auth-adjacent path" +} + +test_brief_only_mode_before_worktree_exists() { + local case_dir out status + case_dir="$TMP_ROOT/brief-only" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf 'Rotate the payment provider credentials.\n' > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "brief-only: brief-only checkpoint must work before any meta/worktree exists" + assert_contains "$out" "RISK:" "brief-only: should still trip on the brief text alone" + pass "fm-risk-tripwire checks the brief alone before a task has been spawned" +} + +test_nothing_to_check_errors() { + local case_dir out status + case_dir="$TMP_ROOT/nothing-to-check" + mkdir -p "$case_dir/state" "$case_dir/data" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1 2>&1) + status=$? + set -e + + expect_code 2 "$status" "nothing-to-check: neither a brief nor meta must error distinctly" + assert_contains "$out" "nothing to check" "nothing-to-check: should explain there was nothing to check" + pass "fm-risk-tripwire errors distinctly when neither a brief nor a usable worktree exists" +} + +test_clean_brief_and_diff_passes +test_brief_keyword_trips_wire +test_diff_path_trips_wire +test_brief_only_mode_before_worktree_exists +test_nothing_to_check_errors + +echo "# all fm-risk-tripwire tests passed" diff --git a/tests/fm-tier-guard.test.sh b/tests/fm-tier-guard.test.sh new file mode 100644 index 000000000..e2000432a --- /dev/null +++ b/tests/fm-tier-guard.test.sh @@ -0,0 +1,192 @@ +#!/usr/bin/env bash +# Tests for bin/fm-tier-guard.sh: the trivial (haiku/low) dispatch tier must +# escalate when its actual diff or elapsed time outgrows its assigned envelope, +# and any tier must escalate once its diff crosses the general heavy-scale +# ceiling, regardless of model/effort. +set -u + +# shellcheck source=tests/lib.sh +. "$(dirname "${BASH_SOURCE[0]}")/lib.sh" +fm_git_identity fmtest fmtest@example.invalid + +TIER_GUARD="$ROOT/bin/fm-tier-guard.sh" +TMP_ROOT=$(fm_test_tmproot fm-tier-guard-tests) + +make_case() { + local name=$1 case_dir + case_dir="$TMP_ROOT/$name" + mkdir -p "$case_dir/state" + + git init -q --bare "$case_dir/origin.git" + git -C "$case_dir/origin.git" symbolic-ref HEAD refs/heads/main + git clone -q "$case_dir/origin.git" "$case_dir/_seed" 2>/dev/null + printf 'base\n' > "$case_dir/_seed/feature.txt" + git -C "$case_dir/_seed" add feature.txt + git -C "$case_dir/_seed" -c user.email=t@t -c user.name=t commit -qm "origin baseline" + git -C "$case_dir/_seed" push -q origin main + rm -rf "$case_dir/_seed" + + git clone -q "$case_dir/origin.git" "$case_dir/project" + git -C "$case_dir/project" remote set-head origin main 2>/dev/null || true + git -C "$case_dir/project" worktree add -q -b "fm/$name" "$case_dir/wt" main + + printf '%s\n' "$case_dir" +} + +write_task_meta() { + local case_dir=$1 id=$2 + shift 2 + fm_write_meta "$case_dir/state/$id.meta" \ + "window=fm-$id" \ + "worktree=$case_dir/wt" \ + "project=$case_dir/project" \ + "$@" +} + +run_tier_guard() { + local case_dir=$1 id=$2 + FM_ROOT_OVERRIDE="$ROOT" \ + FM_STATE_OVERRIDE="$case_dir/state" \ + "$TIER_GUARD" "$id" +} + +test_trivial_tier_within_envelope_passes() { + local case_dir out status + case_dir=$(make_case within-envelope) + printf 'base\nsmall change\n' > "$case_dir/wt/feature.txt" + git -C "$case_dir/wt" add feature.txt + git -C "$case_dir/wt" commit -qm "small edit" + write_task_meta "$case_dir" task-x1 "model=claude-haiku-4-5" "effort=low" + + set +e + out=$(run_tier_guard "$case_dir" task-x1) + status=$? + set -e + + expect_code 0 "$status" "within-envelope: trivial task inside its size/age ceiling must pass" + [ -z "$out" ] || fail "within-envelope: expected no ESCALATE output, got: $out" + pass "fm-tier-guard passes a trivial task within its size and age envelope" +} + +test_trivial_tier_exceeds_line_ceiling_escalates() { + local case_dir out status i + case_dir=$(make_case line-ceiling) + : > "$case_dir/wt/feature.txt" + for i in $(seq 1 40); do echo "line $i" >> "$case_dir/wt/feature.txt"; done + git -C "$case_dir/wt" add feature.txt + git -C "$case_dir/wt" commit -qm "big edit" + write_task_meta "$case_dir" task-x1 "model=claude-haiku-4-5" "effort=low" + + set +e + out=$(run_tier_guard "$case_dir" task-x1) + status=$? + set -e + + expect_code 1 "$status" "line-ceiling: trivial task past the line ceiling must escalate" + assert_contains "$out" "ESCALATE: trivial (haiku/low) task task-x1" "line-ceiling: should name the trivial-tier escalation" + assert_contains "$out" "bump to at least sonnet/high" "line-ceiling: should say what to bump to" + pass "fm-tier-guard escalates a trivial task once its diff exceeds the line ceiling" +} + +test_trivial_tier_exceeds_file_ceiling_escalates() { + local case_dir out status + case_dir=$(make_case file-ceiling) + printf 'a\n' > "$case_dir/wt/a.txt" + printf 'b\n' > "$case_dir/wt/b.txt" + printf 'c\n' > "$case_dir/wt/c.txt" + git -C "$case_dir/wt" add a.txt b.txt c.txt + git -C "$case_dir/wt" commit -qm "three new files" + write_task_meta "$case_dir" task-x1 "model=claude-haiku-4-5" "effort=low" + + set +e + out=$(run_tier_guard "$case_dir" task-x1) + status=$? + set -e + + expect_code 1 "$status" "file-ceiling: trivial task past the file ceiling must escalate" + assert_contains "$out" "ESCALATE: trivial (haiku/low) task task-x1" "file-ceiling: should name the trivial-tier escalation" + pass "fm-tier-guard escalates a trivial task once its diff exceeds the file ceiling" +} + +test_trivial_tier_stale_age_escalates() { + local case_dir out status + case_dir=$(make_case stale-age) + printf 'base\nsmall change\n' > "$case_dir/wt/feature.txt" + git -C "$case_dir/wt" add feature.txt + git -C "$case_dir/wt" commit -qm "small edit" + write_task_meta "$case_dir" task-x1 "model=claude-haiku-4-5" "effort=low" + touch -t 202001010000 "$case_dir/state/task-x1.meta" + + set +e + out=$(run_tier_guard "$case_dir" task-x1) + status=$? + set -e + + expect_code 1 "$status" "stale-age: trivial task past the age ceiling must escalate" + assert_contains "$out" "has run" "stale-age: should name the age escalation" + assert_contains "$out" "bump to at least sonnet/high" "stale-age: should say what to bump to" + pass "fm-tier-guard escalates a trivial task that has run past its age ceiling" +} + +test_non_trivial_tier_small_diff_no_ceiling() { + local case_dir out status i + case_dir=$(make_case non-trivial-ok) + : > "$case_dir/wt/feature.txt" + for i in $(seq 1 40); do echo "line $i" >> "$case_dir/wt/feature.txt"; done + git -C "$case_dir/wt" add feature.txt + git -C "$case_dir/wt" commit -qm "ordinary sonnet-tier edit" + write_task_meta "$case_dir" task-x1 "model=claude-sonnet-5" "effort=high" + + set +e + out=$(run_tier_guard "$case_dir" task-x1) + status=$? + set -e + + expect_code 0 "$status" "non-trivial-ok: sonnet/high has no trivial-tier ceiling to trip" + [ -z "$out" ] || fail "non-trivial-ok: expected no ESCALATE output, got: $out" + pass "fm-tier-guard does not apply the trivial-tier ceiling to a non-trivial task" +} + +test_any_tier_past_heavy_scale_escalates() { + local case_dir out status i + case_dir=$(make_case heavy-scale) + : > "$case_dir/wt/feature.txt" + for i in $(seq 1 6); do echo "line $i" >> "$case_dir/wt/feature.txt"; done + git -C "$case_dir/wt" add feature.txt + git -C "$case_dir/wt" commit -qm "sonnet-tier edit that turned out large" + write_task_meta "$case_dir" task-x1 "model=claude-sonnet-5" "effort=high" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" \ + FM_TIER_HEAVY_MIN_LINES=5 FM_TIER_HEAVY_MIN_FILES=5 "$TIER_GUARD" task-x1) + status=$? + set -e + + expect_code 1 "$status" "heavy-scale: any tier past the heavy-scale ceiling must escalate" + assert_contains "$out" "exceeds the heavy-scale ceiling" "heavy-scale: should name the general heavy-scale escalation" + pass "fm-tier-guard escalates any tier once its diff crosses the general heavy-scale ceiling" +} + +test_missing_meta_errors() { + local case_dir out status + case_dir=$(make_case missing-meta) + + set +e + out=$(run_tier_guard "$case_dir" task-x1 2>&1) + status=$? + set -e + + expect_code 1 "$status" "missing-meta: no meta file must error, not silently pass" + assert_contains "$out" "no meta for task" "missing-meta: should report the missing meta file" + pass "fm-tier-guard errors when the task has no recorded meta" +} + +test_trivial_tier_within_envelope_passes +test_trivial_tier_exceeds_line_ceiling_escalates +test_trivial_tier_exceeds_file_ceiling_escalates +test_trivial_tier_stale_age_escalates +test_non_trivial_tier_small_diff_no_ceiling +test_any_tier_past_heavy_scale_escalates +test_missing_meta_errors + +echo "# all fm-tier-guard tests passed" diff --git a/tests/fm-ultracode-guard.test.sh b/tests/fm-ultracode-guard.test.sh new file mode 100644 index 000000000..ddb4c5cdb --- /dev/null +++ b/tests/fm-ultracode-guard.test.sh @@ -0,0 +1,168 @@ +#!/usr/bin/env bash +# Tests for bin/fm-ultracode-guard.sh: an ultracode-flagged task must not reach +# PR-ready until a genuinely separate, independently-dispatched task recorded +# itself as having reviewed the finished diff - never a self-reference or a +# made-up id. +set -u + +# shellcheck source=tests/lib.sh +. "$(dirname "${BASH_SOURCE[0]}")/lib.sh" + +GUARD="$ROOT/bin/fm-ultracode-guard.sh" +TMP_ROOT=$(fm_test_tmproot fm-ultracode-guard-tests) + +new_state_dir() { + local name=$1 dir + dir="$TMP_ROOT/$name/state" + mkdir -p "$dir" + printf '%s\n' "$dir" +} + +run_guard() { + local state=$1 + shift + FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$state" "$GUARD" "$@" +} + +test_check_passes_when_never_flagged() { + local state status + state=$(new_state_dir never-flagged) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + + set +e + run_guard "$state" check task-x1 >/dev/null 2>&1 + status=$? + set -e + + expect_code 0 "$status" "never-flagged: an unflagged task must pass check" + pass "fm-ultracode-guard check passes a task that was never ultracode-flagged" +} + +test_flag_then_check_fails_until_reviewed() { + local state out status + state=$(new_state_dir flag-then-check) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + + run_guard "$state" flag task-x1 >/dev/null + + set +e + out=$(run_guard "$state" check task-x1 2>&1) + status=$? + set -e + + expect_code 1 "$status" "flag-then-check: a flagged, unreviewed task must fail check" + assert_contains "$out" "is ultracode-flagged" "flag-then-check: should explain why it is blocked" + assert_contains "$out" "role=independent-review" "flag-then-check: should report the default role" + pass "fm-ultracode-guard check refuses a flagged task with no recorded review" +} + +test_flag_custom_role_reported_in_check() { + local state out + state=$(new_state_dir custom-role) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + + run_guard "$state" flag task-x1 breadth-fanout >/dev/null + + set +e + out=$(run_guard "$state" check task-x1 2>&1) + set -e + + assert_contains "$out" "role=breadth-fanout" "custom-role: check should report the custom role" + pass "fm-ultracode-guard flag records a custom role and check reports it" +} + +test_reviewed_by_self_is_refused() { + local state out status + state=$(new_state_dir self-review) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + run_guard "$state" flag task-x1 >/dev/null + + set +e + out=$(run_guard "$state" reviewed task-x1 task-x1 2>&1) + status=$? + set -e + + expect_code 1 "$status" "self-review: a task cannot review itself" + assert_contains "$out" "distinct from task-x1" "self-review: should explain the refusal" + pass "fm-ultracode-guard reviewed refuses a task naming itself as its own reviewer" +} + +test_reviewed_by_unknown_task_is_refused() { + local state out status + state=$(new_state_dir unknown-reviewer) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + run_guard "$state" flag task-x1 >/dev/null + + set +e + out=$(run_guard "$state" reviewed task-x1 made-up-id 2>&1) + status=$? + set -e + + expect_code 1 "$status" "unknown-reviewer: a made-up reviewer id must be refused" + assert_contains "$out" "no recorded state/made-up-id.meta" "unknown-reviewer: should explain the refusal" + pass "fm-ultracode-guard reviewed refuses a reviewer id with no recorded meta" +} + +test_reviewed_by_distinct_dispatched_task_passes_check() { + local state status + state=$(new_state_dir distinct-reviewer) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + fm_write_meta "$state/task-x2.meta" "worktree=/tmp/y" "project=/tmp/y" + run_guard "$state" flag task-x1 >/dev/null + + run_guard "$state" reviewed task-x1 task-x2 >/dev/null + + set +e + run_guard "$state" check task-x1 >/dev/null 2>&1 + status=$? + set -e + + expect_code 0 "$status" "distinct-reviewer: a genuinely separate reviewer must satisfy check" + pass "fm-ultracode-guard check passes once a distinct dispatched task recorded the review" +} + +test_reviewed_without_flag_is_refused() { + local state out status + state=$(new_state_dir no-flag) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + fm_write_meta "$state/task-x2.meta" "worktree=/tmp/y" "project=/tmp/y" + + set +e + out=$(run_guard "$state" reviewed task-x1 task-x2 2>&1) + status=$? + set -e + + expect_code 1 "$status" "no-flag: reviewed must refuse a task that was never flagged" + assert_contains "$out" "not ultracode-flagged" "no-flag: should explain the refusal" + pass "fm-ultracode-guard reviewed refuses to mark an unflagged task as reviewed" +} + +test_reflag_clears_prior_review() { + local state out status + state=$(new_state_dir reflag-clears) + fm_write_meta "$state/task-x1.meta" "worktree=/tmp/x" "project=/tmp/x" + fm_write_meta "$state/task-x2.meta" "worktree=/tmp/y" "project=/tmp/y" + run_guard "$state" flag task-x1 >/dev/null + run_guard "$state" reviewed task-x1 task-x2 >/dev/null + run_guard "$state" flag task-x1 >/dev/null + + set +e + out=$(run_guard "$state" check task-x1 2>&1) + status=$? + set -e + + expect_code 1 "$status" "reflag-clears: re-flagging must clear the prior review" + assert_contains "$out" "is ultracode-flagged" "reflag-clears: should be blocked again" + pass "fm-ultracode-guard re-flagging starts the review requirement over" +} + +test_check_passes_when_never_flagged +test_flag_then_check_fails_until_reviewed +test_flag_custom_role_reported_in_check +test_reviewed_by_self_is_refused +test_reviewed_by_unknown_task_is_refused +test_reviewed_by_distinct_dispatched_task_passes_check +test_reviewed_without_flag_is_refused +test_reflag_clears_prior_review + +echo "# all fm-ultracode-guard tests passed" From 8e3b1662acb64f963c09b7e82727a966bc9dd37a Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 17:56:23 -0400 Subject: [PATCH 13/41] no-mistakes(review): scope risk-tripwire brief scan, narrow bin path, distinct exit codes --- bin/fm-risk-tripwire.sh | 62 ++++++++++++----- bin/fm-tier-guard.sh | 9 ++- bin/fm-ultracode-guard.sh | 4 +- tests/fm-risk-tripwire.test.sh | 119 +++++++++++++++++++++++++++++++++ 4 files changed, 173 insertions(+), 21 deletions(-) diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh index 8883617f9..9fe448b26 100755 --- a/bin/fm-risk-tripwire.sh +++ b/bin/fm-risk-tripwire.sh @@ -14,14 +14,16 @@ # - the worktree's changed file paths vs its project's default branch, if # state/.meta records worktree=/project= (works after spawn, at # Validate time - the second, binding checkpoint against the real diff) -# Errors (exit 2) if neither is available; that means there is nothing to -# check yet. -# -# Prints one "RISK: " line per hit and exits 1 if any fired; exits 0 -# (silent) when neither surface matched. A hit means: floor this task's -# model/effort to the safety-critical profile (opus/xhigh, ultracode -# independent-review) regardless of which rule the natural-language dispatch -# match picked, per AGENTS.md section 4's risk floor. +# Exit codes: +# 0 no risk signal found +# 1 a RISK hit fired (one "RISK: " line per hit is printed) +# 2 could not check: a malformed invocation, or neither a brief nor a usable +# worktree/project exists yet (nothing to check) +# Distinct codes matter so a caller branching on $? cannot mistake a malformed +# invocation for a real risk hit. A hit means: floor this task's model/effort to +# the safety-critical profile (opus/xhigh, ultracode independent-review) +# regardless of which rule the natural-language dispatch match picked, per +# AGENTS.md section 4's risk floor. # # This is a coarse, unpushed-diff-vs-default-branch name-only comparison, not # the PR-aware exact diff bin/fm-review-diff.sh computes - good enough for a @@ -47,19 +49,47 @@ if [ "${1:-}" = "-h" ] || [ "${1:-}" = "--help" ]; then fi ID=${1:-} -[ -n "$ID" ] || { usage; exit 1; } -[ $# -le 1 ] || { usage; exit 1; } +[ -n "$ID" ] || { usage; exit 2; } +[ $# -le 1 ] || { usage; exit 2; } -# Case-insensitive substring match against the brief's prose. Permissive by -# design (a false positive just means "double-check"; a false negative is the -# dangerous direction for a safety floor). +# Case-insensitive, word-bounded match (tolerating plural/verb inflections such +# as migration -> migrations and encrypt -> encryption) against the brief's prose. +# Word boundaries stop substring false positives like "auth" inside +# "authoritative"; inflection tolerance keeps the safety bias intact, since a +# missed plural is a false negative and false negatives are the dangerous +# direction for a safety floor. TEXT_KEYWORDS='auth|authentication|authorization|session|credential|password|secret|token|payment|billing|migration|schema|security|encrypt|decrypt|permission|access.control|data.deletion|bulk.mutation|public.exposure|breaking.change' +TEXT_REGEX="\\<(${TEXT_KEYWORDS})(s|es|ed|ing|ion|ions)?\\>" + +# Scan only the task-specific body of the brief (the # Task section up to the +# next top-level heading), never the fixed scaffold boilerplate bin/fm-brief.sh +# writes into every brief - that boilerplate contains benign words like +# "authoritative" and "future session" that would otherwise trip the wire on +# every task. Falls back to the whole brief when there is no # Task section +# (a non-standard or hand-written brief), keeping the permissive safety bias. +brief_task_body() { + local body + body=$(awk ' + /^# Task[[:space:]]*$/ { intask=1; next } + intask && /^# / { intask=0 } + intask { print } + ' "$1") + if [ -n "$body" ]; then + printf '%s\n' "$body" + else + cat "$1" + fi +} -# Substring match against one changed file's path (already lowercased by the caller). +# Substring match against one changed file's path (already lowercased by the +# caller). Deliberately NOT a blanket bin/* match: firstmate's own supervision +# backbone lives under bin/ and is routed by an explicit dispatch rule (see +# docs/examples/crew-dispatch.json), so flooring every bin/ change here would +# override that rule and defeat the finer tiers. A genuinely risky script still +# trips via its name (e.g. bin/run-migration.sh, bin/auth-setup.sh). path_is_risky() { case "$1" in .github/workflows/*|*/.github/workflows/*) return 0 ;; - bin/*|*/bin/*) return 0 ;; *dockerfile*|*docker-compose*) return 0 ;; *auth*|*migrat*|*schema*|*secret*|*credential*|*payment*|*billing*|*security*|*session*) return 0 ;; esac @@ -72,7 +102,7 @@ CHECKED=0 BRIEF="$DATA/$ID/brief.md" if [ -f "$BRIEF" ]; then CHECKED=1 - hit=$(grep -Eoi "$TEXT_KEYWORDS" "$BRIEF" | tr '[:upper:]' '[:lower:]' | sort -u | tr '\n' ',' | sed 's/,$//') + hit=$(brief_task_body "$BRIEF" | grep -Eoi "$TEXT_REGEX" | tr '[:upper:]' '[:lower:]' | sort -u | tr '\n' ',' | sed 's/,$//') if [ -n "$hit" ]; then echo "RISK: brief for $ID mentions risk-adjacent term(s): $hit" FOUND=1 diff --git a/bin/fm-tier-guard.sh b/bin/fm-tier-guard.sh index be7effca2..ada2385db 100755 --- a/bin/fm-tier-guard.sh +++ b/bin/fm-tier-guard.sh @@ -9,8 +9,13 @@ # Usage: fm-tier-guard.sh # # Reads state/.meta for model=/effort=, and that file's own mtime as a -# spawn-time proxy. Reuses bin/fm-review-diff.sh --stat for the actual diff size -# instead of re-deriving the authoritative base/PR-head resolution here. +# best-effort spawn-time proxy for the age check - NOT authoritative. A later +# append to meta (bin/fm-pr-check.sh recording pr=/pr_head=, or fm-x-link +# recording X-mode fields) resets the mtime, so a long-running task can read as +# younger than it is and slip the age ceiling; the diff size/file ceilings are +# the primary signal and are unaffected by that reset. Reuses +# bin/fm-review-diff.sh --stat for the actual diff size instead of re-deriving +# the authoritative base/PR-head resolution here. # # Prints one "ESCALATE: " line per triggered condition and exits 1 if # any fired; exits 0 (silent) when the task is within its tier's envelope. diff --git a/bin/fm-ultracode-guard.sh b/bin/fm-ultracode-guard.sh index 099ae8edc..ed0de277a 100755 --- a/bin/fm-ultracode-guard.sh +++ b/bin/fm-ultracode-guard.sh @@ -56,9 +56,7 @@ cmd_flag() { # Overwrites any existing marker: re-flagging (e.g. after an escalation) # deliberately clears a prior reviewed_by=, since that review was against an # earlier version of the diff and the requirement starts over. - { - echo "role=$role" - } > "$MARKER" + echo "role=$role" > "$MARKER" echo "flagged $ID ultracode role=$role" } diff --git a/tests/fm-risk-tripwire.test.sh b/tests/fm-risk-tripwire.test.sh index fc4999fdc..edfc58140 100644 --- a/tests/fm-risk-tripwire.test.sh +++ b/tests/fm-risk-tripwire.test.sh @@ -49,6 +49,18 @@ run_tripwire() { "$TRIPWIRE" task-x1 } +# Scaffold a real ship brief via bin/fm-brief.sh, then substitute the {TASK} +# placeholder with the given task text - exercising the actual scaffold +# boilerplate rather than a hand-written stand-in. +scaffold_brief() { + local case_dir=$1 task=$2 brief + mkdir -p "$case_dir/state" + FM_ROOT_OVERRIDE="$ROOT" FM_DATA_OVERRIDE="$case_dir/data" FM_STATE_OVERRIDE="$case_dir/state" \ + "$ROOT/bin/fm-brief.sh" task-x1 someproject >/dev/null 2>&1 + brief="$case_dir/data/task-x1/brief.md" + sed "s|{TASK}|$task|" "$brief" > "$brief.tmp" && mv "$brief.tmp" "$brief" +} + test_clean_brief_and_diff_passes() { local case_dir out status case_dir=$(make_case clean) @@ -137,10 +149,117 @@ test_nothing_to_check_errors() { pass "fm-risk-tripwire errors distinctly when neither a brief nor a usable worktree exists" } +test_scaffolded_brief_boilerplate_does_not_trip() { + local case_dir out status + case_dir="$TMP_ROOT/scaffold-clean" + scaffold_brief "$case_dir" "Fix a typo in the CLI help text." + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1 2>&1) + status=$? + set -e + + expect_code 0 "$status" "scaffold-clean: a real scaffolded ship brief with a benign task must not trip" + [ -z "$out" ] || fail "scaffold-clean: expected no RISK output from scaffold boilerplate, got: $out" + pass "fm-risk-tripwire does not trip on fm-brief.sh scaffold boilerplate" +} + +test_scaffolded_brief_risky_task_still_trips() { + local case_dir out status + case_dir="$TMP_ROOT/scaffold-risky" + scaffold_brief "$case_dir" "Rotate the payment credentials and run the schema migration." + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "scaffold-risky: a risk-worded task body inside a real scaffold must still trip" + assert_contains "$out" "RISK: brief for task-x1" "scaffold-risky: should name the brief hit" + pass "fm-risk-tripwire still scans the task body of a scaffolded brief" +} + +test_word_boundary_avoids_substring_false_positive() { + local case_dir out status + case_dir="$TMP_ROOT/word-boundary" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nMake the config loader the authoritative source of truth.\n\n# Setup\nnothing risky here.\n' \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 0 "$status" "word-boundary: 'authoritative' must not match the 'auth' keyword" + [ -z "$out" ] || fail "word-boundary: expected no RISK output, got: $out" + pass "fm-risk-tripwire does not treat 'authoritative' as an auth keyword hit" +} + +test_inflected_keyword_still_trips() { + local case_dir out status + case_dir="$TMP_ROOT/inflected" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nRun the pending database migrations and rotate the tokens.\n\n# Setup\nx\n' \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "inflected: a plural risk word must still trip (no false negative)" + assert_contains "$out" "migration" "inflected: should surface the matched migration term" + pass "fm-risk-tripwire still trips on inflected/plural risk words" +} + +test_supervision_bin_path_does_not_trip() { + local case_dir out status + case_dir=$(make_case bin-path) + printf 'Tune the watcher poll cadence.\n' > "$case_dir/data/task-x1/brief.md" + mkdir -p "$case_dir/wt/bin" + printf 'watcher tweak\n' > "$case_dir/wt/bin/fm-watch.sh" + printf 'guard tweak\n' > "$case_dir/wt/bin/fm-guard.sh" + git -C "$case_dir/wt" add bin/fm-watch.sh bin/fm-guard.sh + git -C "$case_dir/wt" commit -qm "tweak supervision backbone" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 0 "$status" "bin-path: a supervision-backbone bin/ change alone must not trip the wire" + [ -z "$out" ] || fail "bin-path: expected no RISK output, got: $out" + pass "fm-risk-tripwire does not trip on a supervision-backbone bin/ path" +} + +test_usage_error_exit_code() { + local status + set +e + FM_ROOT_OVERRIDE="$ROOT" "$TRIPWIRE" >/dev/null 2>&1 + status=$? + set -e + expect_code 2 "$status" "usage-empty-id: a malformed invocation must exit 2, not 1 (the RISK code)" + + set +e + FM_ROOT_OVERRIDE="$ROOT" "$TRIPWIRE" one two >/dev/null 2>&1 + status=$? + set -e + expect_code 2 "$status" "usage-extra-args: extra args must exit 2, not 1 (the RISK code)" + pass "fm-risk-tripwire uses a distinct exit code for malformed invocations" +} + test_clean_brief_and_diff_passes test_brief_keyword_trips_wire test_diff_path_trips_wire test_brief_only_mode_before_worktree_exists test_nothing_to_check_errors +test_scaffolded_brief_boilerplate_does_not_trip +test_scaffolded_brief_risky_task_still_trips +test_word_boundary_avoids_substring_false_positive +test_inflected_keyword_still_trips +test_supervision_bin_path_does_not_trip +test_usage_error_exit_code echo "# all fm-risk-tripwire tests passed" From aec0b3c5b9d7f3f7a5ea7ee0514cd6f36c1c61c3 Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 19:14:11 -0400 Subject: [PATCH 14/41] no-mistakes(review): harden risk-tripwire scans and mechanically wire ultracode floor --- AGENTS.md | 1 + bin/fm-risk-tripwire.sh | 101 +++++++++--- tests/fm-risk-tripwire.test.sh | 270 +++++++++++++++++++++++++++++++++ 3 files changed, 349 insertions(+), 23 deletions(-) diff --git a/AGENTS.md b/AGENTS.md index 535398658..656a37bcb 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -88,6 +88,7 @@ When a resolved profile sets `ultracode` (schema in `docs/configuration.md` "Cre **Risk floor.** Run `bin/fm-risk-tripwire.sh ` before spawn and again once the task has a brief or a diff. A hit floors the model/effort to the safety-critical profile (`opus`/`xhigh`, ultracode `independent-review`) regardless of which rule matched, and a hit found only at Validate time still floors the task in place and is never silently downgraded afterward. +On a hit, also run `bin/fm-ultracode-guard.sh flag independent-review` so the independent-review requirement is tracked mechanically even when the matched rule did not set ultracode. Before a dispatch rule may name a specific non-default model, verify it end-to-end on a trivial supervised task, exactly as this section already requires before naming a new harness adapter; do not use an unverified model live. diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh index 9fe448b26..5d0c17d85 100755 --- a/bin/fm-risk-tripwire.sh +++ b/bin/fm-risk-tripwire.sh @@ -52,27 +52,48 @@ ID=${1:-} [ -n "$ID" ] || { usage; exit 2; } [ $# -le 1 ] || { usage; exit 2; } -# Case-insensitive, word-bounded match (tolerating plural/verb inflections such -# as migration -> migrations and encrypt -> encryption) against the brief's prose. -# Word boundaries stop substring false positives like "auth" inside -# "authoritative"; inflection tolerance keeps the safety bias intact, since a -# missed plural is a false negative and false negatives are the dangerous -# direction for a safety floor. -TEXT_KEYWORDS='auth|authentication|authorization|session|credential|password|secret|token|payment|billing|migration|schema|security|encrypt|decrypt|permission|access.control|data.deletion|bulk.mutation|public.exposure|breaking.change' +# Case-insensitive, word-bounded match against the brief's prose, tolerating the +# inflections and affixes real task phrasing uses. Word boundaries stop substring +# false positives like "auth" inside "authoritative", but three things must still +# reach the scan or a genuinely risky task slips to a cheap tier (the dangerous +# direction for a safety floor): +# - verb forms need their own stems: appending the outer (s|ed|ing|ion|...) +# suffix to a bare NOUN literal cannot reach the verb, so "migration" alone +# could only become "migrations", never "migrate"/"migrating"/"migrated", +# and the bare "auth" branch never reaches "authorize"/"authenticate". +# - the un/re/de/pre prefixes ("unauthorized"/"unauthenticated" - the most +# common access-control phrasing) glue a letter onto the stem's front and +# defeat the leading \<, so the auth stems carry an explicit optional prefix. +# - a bare auth\w* stem is still avoided, so "authoritative"/"author" stay shut. +# Snake_case identifiers in the body are split on '_' before matching (the '_' is +# a regex word char, so "run_schema_migration" would otherwise hide its words). +TEXT_KEYWORDS='auth|authentication|authorization|(un|re|de|pre)?authoriz(e|ed|es|ing|ation|ations|er|ers)?|(un|re|de|pre)?authenticat(e|ed|es|ing|ion|ions|or|ors)?|session|credential|password|secret|token|payment|billing|migrat(e|ed|es|ing|ion|ions)|schema|security|encrypt|decrypt|permission|access.control|data.deletion|bulk.mutation|public.exposure|breaking.change' TEXT_REGEX="\\<(${TEXT_KEYWORDS})(s|es|ed|ing|ion|ions)?\\>" # Scan only the task-specific body of the brief (the # Task section up to the -# next top-level heading), never the fixed scaffold boilerplate bin/fm-brief.sh -# writes into every brief - that boilerplate contains benign words like -# "authoritative" and "future session" that would otherwise trip the wire on -# every task. Falls back to the whole brief when there is no # Task section -# (a non-standard or hand-written brief), keeping the permissive safety bias. +# next scaffold section heading), never the fixed scaffold boilerplate +# bin/fm-brief.sh writes into every brief - that boilerplate contains benign +# words like "authoritative" and "future session" that would otherwise trip the +# wire on every task. The section boundary is one of the scaffold's own known +# headings (# Setup/# Rules/# Project memory/# Definition of done), not any +# column-0 "# " line: a shell/code comment embedded in the task body (e.g. +# "# then run the schema migration") must NOT end the scan, or the risk words +# after it are silently dropped - the dangerous direction for a safety floor. +# A boundary heading must also be blank-line-preceded, as every real scaffold +# heading is (bin/fm-brief.sh emits a blank line before each), so a task body +# that itself quotes a bare "# Setup" line inline does not cut the scan short. +# Lines inside a fenced code block are likewise never treated as a boundary. +# Falls back to the whole brief when there is no # Task section (a non-standard +# or hand-written brief), keeping the permissive safety bias. brief_task_body() { local body body=$(awk ' - /^# Task[[:space:]]*$/ { intask=1; next } - intask && /^# / { intask=0 } + { blank = ($0 ~ /^[[:space:]]*$/) } + /^```/ { fence = !fence; prevblank = blank; next } + !fence && /^# Task[[:space:]]*$/ { intask=1; prevblank = blank; next } + intask && !fence && prevblank && /^# (Setup|Rules|Project memory|Definition of done)[[:space:]]*$/ { intask=0 } intask { print } + { prevblank = blank } ' "$1") if [ -n "$body" ]; then printf '%s\n' "$body" @@ -81,18 +102,46 @@ brief_task_body() { fi } -# Substring match against one changed file's path (already lowercased by the -# caller). Deliberately NOT a blanket bin/* match: firstmate's own supervision -# backbone lives under bin/ and is routed by an explicit dispatch rule (see -# docs/examples/crew-dispatch.json), so flooring every bin/ change here would -# override that rule and defeat the finer tiers. A genuinely risky script still -# trips via its name (e.g. bin/run-migration.sh, bin/auth-setup.sh). +# Component/token match against one changed file's path (already lowercased by +# the caller). Deliberately NOT a blanket bin/* match: firstmate's own +# supervision backbone lives under bin/ and is routed by an explicit dispatch +# rule (see docs/examples/crew-dispatch.json), so flooring every bin/ change +# here would override that rule and defeat the finer tiers. +# +# The same word-boundary discipline as the brief-text scan applies here, so a +# risk word must be a real path component or delimited token, not a bare +# substring: "strong" words (auth/migrat/schema/... families) match anywhere +# they appear as a /, -, _, or . delimited token, catching bin/auth-setup.sh, +# bin/run-migration.sh, and db/schema.sql; the weaker "session" matches only as +# a whole path component or filename base (lib/auth/session.rb, +# app/models/session.rb), never as a hyphen fragment - otherwise the +# supervision backbone bin/fm-session-start.sh would over-match. Bare-substring +# false positives like "auth" inside AUTHORS/docs/authors.md no longer trip. +# The flip side is intentional: a risk word glued into a compound component with +# NO delimiter (e.g. "authsetup.rb") is not matched, because catching it would +# require substring matching again and reopen exactly those false positives; the +# brief-text scan and the delimiter tokenization below cover the realistic cases. +PATH_STRONG_REGEX='^(auth|authn|authz|authoriz(e|ed|es|ing|ation|ations|er|ers)?|authenticat(e|ed|es|ing|ion|ions|or|ors)?|migrat(e|ed|es|ing|ion|ions)|schema|schemas|secret|secrets|credential|credentials|payment|payments|billing|security)$' +PATH_WEAK_REGEX='^(session|sessions)$' path_is_risky() { - case "$1" in + local path=$1 comp base + case "$path" in .github/workflows/*|*/.github/workflows/*) return 0 ;; *dockerfile*|*docker-compose*) return 0 ;; - *auth*|*migrat*|*schema*|*secret*|*credential*|*payment*|*billing*|*security*|*session*) return 0 ;; esac + while IFS= read -r comp; do + [ -n "$comp" ] || continue + base=${comp%.*} + # Whole path component or filename base equal to a strong or weak word. + if printf '%s\n%s\n' "$comp" "$base" | grep -Eiq "$PATH_STRONG_REGEX|$PATH_WEAK_REGEX"; then + return 0 + fi + # Strong words also match as a ., -, or _ delimited token inside a compound + # name (e.g. config.schema.json, run-migration.sh), the fail-safe direction. + if printf '%s\n' "$base" | tr '._-' '\n' | grep -Eiq "$PATH_STRONG_REGEX"; then + return 0 + fi + done < <(printf '%s\n' "$path" | tr '/' '\n') return 1 } @@ -102,7 +151,7 @@ CHECKED=0 BRIEF="$DATA/$ID/brief.md" if [ -f "$BRIEF" ]; then CHECKED=1 - hit=$(brief_task_body "$BRIEF" | grep -Eoi "$TEXT_REGEX" | tr '[:upper:]' '[:lower:]' | sort -u | tr '\n' ',' | sed 's/,$//') + hit=$(brief_task_body "$BRIEF" | tr '_' ' ' | grep -Eoi "$TEXT_REGEX" | tr '[:upper:]' '[:lower:]' | sort -u | tr '\n' ',' | sed 's/,$//') if [ -n "$hit" ]; then echo "RISK: brief for $ID mentions risk-adjacent term(s): $hit" FOUND=1 @@ -115,6 +164,12 @@ if [ -f "$META" ]; then PROJ=$(grep '^project=' "$META" | tail -1 | cut -d= -f2- || true) if [ -n "$WT" ] && [ -n "$PROJ" ] && [ -d "$WT" ] && [ -d "$PROJ" ]; then CHECKED=1 + # The base resolution below duplicates bin/fm-review-diff.sh's origin fetch + # and origin/ verification on purpose: this scan needs the changed + # PATH LIST, but fm-review-diff.sh only exposes --stat (a diffstat), not a + # --name-only path list, so there is nothing to reuse for a name scan. + # Adding a --name-only mode to that script is the clean fix and is left as a + # follow-up rather than expanding this task's scope into another owner. DEFAULT=$(fm_default_branch "$PROJ" 2>/dev/null || true) if [ -n "$DEFAULT" ]; then BASE="$DEFAULT" diff --git a/tests/fm-risk-tripwire.test.sh b/tests/fm-risk-tripwire.test.sh index edfc58140..a651fa55d 100644 --- a/tests/fm-risk-tripwire.test.sh +++ b/tests/fm-risk-tripwire.test.sh @@ -250,6 +250,263 @@ test_usage_error_exit_code() { pass "fm-risk-tripwire uses a distinct exit code for malformed invocations" } +test_embedded_comment_task_body_still_scanned() { + local case_dir out status + case_dir="$TMP_ROOT/embedded-comment" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + # A column-0 "# " line inside the task body (a shell comment in an example + # command) must NOT end the Task-section scan, or the risk words after it are + # silently dropped - the dangerous direction for a safety floor. + printf '# Task\nImplement the DB runner. Example invocation:\n# then run the schema migration and rotate the tokens\n./run up\n\n# Setup\nnothing risky here.\n' \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "embedded-comment: risk text after an embedded '# ' line must still be scanned" + assert_contains "$out" "schema" "embedded-comment: should surface the term on the embedded comment line" + assert_contains "$out" "migration" "embedded-comment: should surface the migration term after the comment line" + pass "fm-risk-tripwire keeps scanning the task body past an embedded '# ' comment line" +} + +test_auth_verbs_trip_wire() { + local verb case_dir out status i=0 + for verb in authorize authorized authorizing authenticate authenticated; do + i=$((i + 1)) + case_dir="$TMP_ROOT/auth-verb-$i" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nAdd middleware to %s admin requests.\n\n# Setup\nx\n' "$verb" \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "auth-verb: '$verb' must trip the wire" + assert_contains "$out" "RISK: brief for task-x1" "auth-verb: '$verb' should name the brief hit" + done + pass "fm-risk-tripwire trips on auth verbs (authorize/authenticate families)" +} + +test_auth_nouns_do_not_false_positive() { + local word case_dir out status i=0 + for word in authoritative author; do + i=$((i + 1)) + case_dir="$TMP_ROOT/auth-noun-$i" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nMake the loader the %s source of config.\n\n# Setup\nx\n' "$word" \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 0 "$status" "auth-noun: '$word' must not trip the wire" + [ -z "$out" ] || fail "auth-noun: '$word' expected no RISK output, got: $out" + done + pass "fm-risk-tripwire does not treat 'authoritative'/'author' as auth hits" +} + +test_session_start_bin_path_does_not_trip() { + local case_dir out status + case_dir=$(make_case session-start) + printf 'Tune the digest ordering.\n' > "$case_dir/data/task-x1/brief.md" + mkdir -p "$case_dir/wt/bin" + printf 'digest tweak\n' > "$case_dir/wt/bin/fm-session-start.sh" + git -C "$case_dir/wt" add bin/fm-session-start.sh + git -C "$case_dir/wt" commit -qm "tweak session-start" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 0 "$status" "session-start: 'session' as a hyphen fragment of a supervision script must not trip" + [ -z "$out" ] || fail "session-start: expected no RISK output, got: $out" + pass "fm-risk-tripwire does not trip on bin/fm-session-start.sh" +} + +test_auth_setup_bin_path_trips() { + local case_dir out status + case_dir=$(make_case auth-setup) + printf 'Wire up the setup helper.\n' > "$case_dir/data/task-x1/brief.md" + mkdir -p "$case_dir/wt/bin" + printf 'setup\n' > "$case_dir/wt/bin/auth-setup.sh" + git -C "$case_dir/wt" add bin/auth-setup.sh + git -C "$case_dir/wt" commit -qm "add auth setup" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 1 "$status" "auth-setup: 'auth' as a real hyphen token must still trip under bin/" + assert_contains "$out" "bin/auth-setup.sh" "auth-setup: should list the risky path" + pass "fm-risk-tripwire still trips on bin/auth-setup.sh via its auth token" +} + +test_dot_delimited_strong_token_trips() { + local case_dir out status + case_dir=$(make_case dot-token) + printf 'Update the generated config file.\n' > "$case_dir/data/task-x1/brief.md" + mkdir -p "$case_dir/wt/config" + printf 'x\n' > "$case_dir/wt/config/db.schema.json" + git -C "$case_dir/wt" add config/db.schema.json + git -C "$case_dir/wt" commit -qm "add db schema config" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 1 "$status" "dot-token: 'schema' as an interior dot token must trip" + assert_contains "$out" "config/db.schema.json" "dot-token: should list the risky path" + pass "fm-risk-tripwire trips on a strong risk word as a dot-delimited token" +} + +test_authors_doc_path_does_not_trip() { + local case_dir out status + case_dir=$(make_case authors-doc) + printf 'Add a contributors list.\n' > "$case_dir/data/task-x1/brief.md" + mkdir -p "$case_dir/wt/docs" + printf 'names\n' > "$case_dir/wt/docs/authors.md" + git -C "$case_dir/wt" add docs/authors.md + git -C "$case_dir/wt" commit -qm "add authors doc" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 0 "$status" "authors-doc: 'authors' is not the 'auth' token, must not trip" + [ -z "$out" ] || fail "authors-doc: expected no RISK output, got: $out" + pass "fm-risk-tripwire does not trip on docs/authors.md" +} + +test_migrate_verbs_trip_wire() { + local verb case_dir out status i=0 + for verb in migrate migrating migrated; do + i=$((i + 1)) + case_dir="$TMP_ROOT/migrate-verb-$i" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\n%s the customers table to the new engine.\n\n# Setup\nx\n' "$verb" \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "migrate-verb: '$verb' must trip the wire" + assert_contains "$out" "RISK: brief for task-x1" "migrate-verb: '$verb' should name the brief hit" + done + pass "fm-risk-tripwire trips on migrate verb forms (migrate/migrating/migrated)" +} + +test_auth_prefix_forms_trip_wire() { + local word case_dir out status i=0 + for word in unauthorized unauthenticated reauthenticate deauthorize; do + i=$((i + 1)) + case_dir="$TMP_ROOT/auth-prefix-$i" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nReject %s requests at the gateway.\n\n# Setup\nx\n' "$word" \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "auth-prefix: '$word' must trip the wire" + done + pass "fm-risk-tripwire trips on prefixed auth forms (unauthorized/unauthenticated/...)" +} + +test_authenticator_noun_trips_wire() { + local word case_dir out status i=0 + for word in authenticator authenticators; do + i=$((i + 1)) + case_dir="$TMP_ROOT/authenticator-$i" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nAdd support for hardware %s at login.\n\n# Setup\nx\n' "$word" \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "authenticator: '$word' must trip the wire" + done + pass "fm-risk-tripwire trips on authenticator/authenticators nouns" +} + +test_snake_case_risk_word_trips() { + local case_dir out status + case_dir="$TMP_ROOT/snake-case" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nImplement the runner. Call run_schema_migration_now to apply it.\n\n# Setup\nx\n' \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "snake-case: a risk word inside a snake_case identifier must trip" + assert_contains "$out" "schema" "snake-case: should surface the schema token" + assert_contains "$out" "migration" "snake-case: should surface the migration token" + pass "fm-risk-tripwire splits snake_case identifiers so embedded risk words trip" +} + +test_task_body_inline_heading_still_scanned() { + local case_dir out status + case_dir="$TMP_ROOT/inline-heading" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + # A bare "# Setup" quoted inline in the task body (not blank-line-preceded, as + # the real scaffold heading always is) must NOT terminate the scan early. + printf '# Task\nDo the thing. Configuration follows:\n# Setup\nRotate the credentials and run the migration.\n\n# Setup\nbenign boilerplate goes here.\n' \ + > "$case_dir/data/task-x1/brief.md" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "inline-heading: risk text after an inline (non-blank-preceded) '# Setup' must still be scanned" + assert_contains "$out" "credential" "inline-heading: should surface the credentials term" + assert_contains "$out" "migration" "inline-heading: should surface the migration term" + pass "fm-risk-tripwire keeps scanning past an inline heading that is not the scaffold boundary" +} + +test_authorizer_path_trips() { + local case_dir out status + case_dir=$(make_case authorizer-path) + printf 'Refactor the request pipeline.\n' > "$case_dir/data/task-x1/brief.md" + mkdir -p "$case_dir/wt/app/authorizers" + printf 'x\n' > "$case_dir/wt/app/authorizers/user_authorizer.rb" + git -C "$case_dir/wt" add app/authorizers/user_authorizer.rb + git -C "$case_dir/wt" commit -qm "add user authorizer" + write_task_meta "$case_dir" + + set +e + out=$(run_tripwire "$case_dir") + status=$? + set -e + + expect_code 1 "$status" "authorizer-path: an 'authorizer' component must trip (path authoriz stem parity)" + assert_contains "$out" "app/authorizers/user_authorizer.rb" "authorizer-path: should list the risky path" + pass "fm-risk-tripwire trips on authorizer paths" +} + test_clean_brief_and_diff_passes test_brief_keyword_trips_wire test_diff_path_trips_wire @@ -261,5 +518,18 @@ test_word_boundary_avoids_substring_false_positive test_inflected_keyword_still_trips test_supervision_bin_path_does_not_trip test_usage_error_exit_code +test_embedded_comment_task_body_still_scanned +test_auth_verbs_trip_wire +test_auth_nouns_do_not_false_positive +test_session_start_bin_path_does_not_trip +test_auth_setup_bin_path_trips +test_dot_delimited_strong_token_trips +test_authors_doc_path_does_not_trip +test_migrate_verbs_trip_wire +test_auth_prefix_forms_trip_wire +test_authenticator_noun_trips_wire +test_snake_case_risk_word_trips +test_task_body_inline_heading_still_scanned +test_authorizer_path_trips echo "# all fm-risk-tripwire tests passed" From 6c517c96a19470921dbdc6ab8bd4b807abfa8fe1 Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 19:50:32 -0400 Subject: [PATCH 15/41] no-mistakes(review): make risk-tripwire grep portable, give tier-guard distinct exit codes --- bin/fm-risk-tripwire.sh | 33 ++++++++++++----- bin/fm-tier-guard.sh | 16 +++++--- tests/fm-risk-tripwire.test.sh | 67 ++++++++++++++++++++++++++++++++++ tests/fm-tier-guard.test.sh | 21 ++++++++++- 4 files changed, 120 insertions(+), 17 deletions(-) diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh index 5d0c17d85..3e8f1b585 100755 --- a/bin/fm-risk-tripwire.sh +++ b/bin/fm-risk-tripwire.sh @@ -53,7 +53,7 @@ ID=${1:-} [ $# -le 1 ] || { usage; exit 2; } # Case-insensitive, word-bounded match against the brief's prose, tolerating the -# inflections and affixes real task phrasing uses. Word boundaries stop substring +# inflections and affixes real task phrasing uses. Word bounding stops substring # false positives like "auth" inside "authoritative", but three things must still # reach the scan or a genuinely risky task slips to a cheap tier (the dangerous # direction for a safety floor): @@ -62,13 +62,22 @@ ID=${1:-} # could only become "migrations", never "migrate"/"migrating"/"migrated", # and the bare "auth" branch never reaches "authorize"/"authenticate". # - the un/re/de/pre prefixes ("unauthorized"/"unauthenticated" - the most -# common access-control phrasing) glue a letter onto the stem's front and -# defeat the leading \<, so the auth stems carry an explicit optional prefix. -# - a bare auth\w* stem is still avoided, so "authoritative"/"author" stay shut. -# Snake_case identifiers in the body are split on '_' before matching (the '_' is -# a regex word char, so "run_schema_migration" would otherwise hide its words). -TEXT_KEYWORDS='auth|authentication|authorization|(un|re|de|pre)?authoriz(e|ed|es|ing|ation|ations|er|ers)?|(un|re|de|pre)?authenticat(e|ed|es|ing|ion|ions|or|ors)?|session|credential|password|secret|token|payment|billing|migrat(e|ed|es|ing|ion|ions)|schema|security|encrypt|decrypt|permission|access.control|data.deletion|bulk.mutation|public.exposure|breaking.change' -TEXT_REGEX="\\<(${TEXT_KEYWORDS})(s|es|ed|ing|ion|ions)?\\>" +# common access-control phrasing) glue a letter onto the stem's front, so +# the auth stems carry an explicit optional prefix. +# - a bare "auth"-anything stem is still avoided, so "authoritative"/"author" +# stay shut. +# Word bounding is done by splitting the body into whole tokens (below), NOT by +# grep's \< / \> anchors: those are a GNU extension BSD grep does not honor, so +# on macOS they would silently match nothing and let every risky brief through - +# a false negative in the dangerous direction. Whole-token matching is identical +# on GNU and BSD grep and, unlike a boundary-consuming grep -o pattern, still +# reports BOTH words of an adjacent risk pair like "session token" (a consumed +# shared delimiter would drop the second). Single-word keywords go in WORD_REGEX; +# the multi-word phrases have no lone token and are matched on the +# space-normalized stream via PHRASE_REGEX. +WORD_KEYWORDS='auth|authentication|authorization|(un|re|de|pre)?authoriz(e|ed|es|ing|ation|ations|er|ers)?|(un|re|de|pre)?authenticat(e|ed|es|ing|ion|ions|or|ors)?|session|credential|password|secret|token|payment|billing|migrat(e|ed|es|ing|ion|ions)|schema|security|encrypt|decrypt|permission' +WORD_REGEX="(${WORD_KEYWORDS})(s|es|ed|ing|ion|ions)?" +PHRASE_REGEX='(access[[:space:]]+control|data[[:space:]]+deletion|bulk[[:space:]]+mutation|public[[:space:]]+exposure|breaking[[:space:]]+change)(s|es|ed|ing|ion|ions)?' # Scan only the task-specific body of the brief (the # Task section up to the # next scaffold section heading), never the fixed scaffold boilerplate @@ -151,7 +160,13 @@ CHECKED=0 BRIEF="$DATA/$ID/brief.md" if [ -f "$BRIEF" ]; then CHECKED=1 - hit=$(brief_task_body "$BRIEF" | tr '_' ' ' | grep -Eoi "$TEXT_REGEX" | tr '[:upper:]' '[:lower:]' | sort -u | tr '\n' ',' | sed 's/,$//') + lc=$(brief_task_body "$BRIEF" | tr '[:upper:]' '[:lower:]') + # Split on every non-alnum char so each word is its own token (this also + # splits snake_case identifiers like run_schema_migration), then match whole + # tokens; phrases keep their inter-word gap on the space-normalized stream. + word_hits=$(printf '%s\n' "$lc" | tr -c '[:alnum:]' '\n' | grep -xE "$WORD_REGEX" || true) + phrase_hits=$(printf '%s\n' "$lc" | tr -c '[:alnum:]' ' ' | tr -s ' ' | grep -oE "$PHRASE_REGEX" || true) + hit=$(printf '%s\n%s\n' "$word_hits" "$phrase_hits" | sed '/^$/d' | sort -u | tr '\n' ',' | sed 's/,$//') if [ -n "$hit" ]; then echo "RISK: brief for $ID mentions risk-adjacent term(s): $hit" FOUND=1 diff --git a/bin/fm-tier-guard.sh b/bin/fm-tier-guard.sh index ada2385db..38234c4b0 100755 --- a/bin/fm-tier-guard.sh +++ b/bin/fm-tier-guard.sh @@ -17,8 +17,12 @@ # bin/fm-review-diff.sh --stat for the actual diff size instead of re-deriving # the authoritative base/PR-head resolution here. # -# Prints one "ESCALATE: " line per triggered condition and exits 1 if -# any fired; exits 0 (silent) when the task is within its tier's envelope. +# Prints one "ESCALATE: " line per triggered condition. Exit codes: +# 0 within the tier's envelope (silent) +# 1 at least one escalation fired +# 2 usage or setup error (bad args, missing meta, or fm-review-diff failed) +# Distinct codes let a caller branching on $? tell a real escalation from a +# malformed invocation, matching the sibling guardrail bin/fm-risk-tripwire.sh. # Only the trivial (Haiku/low) tier has a size/age ceiling checked here; every # other tier is covered only by the general, tier-independent ceiling below. # On an escalation, bump the task's model/effort in place per AGENTS.md @@ -55,11 +59,11 @@ if [ "${1:-}" = "-h" ] || [ "${1:-}" = "--help" ]; then fi ID=${1:-} -[ -n "$ID" ] || { usage; exit 1; } -[ $# -le 1 ] || { usage; exit 1; } +[ -n "$ID" ] || { usage; exit 2; } +[ $# -le 1 ] || { usage; exit 2; } META="$STATE/$ID.meta" -[ -f "$META" ] || { echo "error: no meta for task $ID at $META" >&2; exit 1; } +[ -f "$META" ] || { echo "error: no meta for task $ID at $META" >&2; exit 2; } MODEL=$(grep '^model=' "$META" | tail -1 | cut -d= -f2- || true) EFFORT=$(grep '^effort=' "$META" | tail -1 | cut -d= -f2- || true) @@ -85,7 +89,7 @@ is_trivial_tier() { STAT_OUT=$("$SCRIPT_DIR/fm-review-diff.sh" "$ID" --stat 2>/dev/null) || { echo "error: fm-review-diff.sh failed for $ID" >&2 - exit 1 + exit 2 } FILES=0 diff --git a/tests/fm-risk-tripwire.test.sh b/tests/fm-risk-tripwire.test.sh index a651fa55d..b29d0ac2a 100644 --- a/tests/fm-risk-tripwire.test.sh +++ b/tests/fm-risk-tripwire.test.sh @@ -507,7 +507,74 @@ test_authorizer_path_trips() { pass "fm-risk-tripwire trips on authorizer paths" } +test_bare_auth_matches_but_authoritative_does_not() { + # Guards the portable word boundary against silently no-op'ing on BSD grep: a + # no-op that matches nothing would miss the bare 'auth' (part b), and a no-op + # that matches substrings would trip on 'authoritative' (part a). Both asserted. + local case_dir out status + # (a) 'authoritative' alone must not match the 'auth' keyword. + case_dir="$TMP_ROOT/authoritative-only" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nMake the loader the authoritative config source only.\n\n# Setup\nx\n' \ + > "$case_dir/data/task-x1/brief.md" + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + expect_code 0 "$status" "authoritative-only: 'authoritative' must not match the auth keyword" + [ -z "$out" ] || fail "authoritative-only: expected no RISK output, got: $out" + + # (b) a bare 'auth' word appearing mid-line (not at the string edges) must trip. + case_dir="$TMP_ROOT/bare-auth-midline" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nPlease auth the request before the handler runs.\n\n# Setup\nx\n' \ + > "$case_dir/data/task-x1/brief.md" + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + expect_code 1 "$status" "bare-auth-midline: a standalone mid-line 'auth' must trip the wire" + assert_contains "$out" "auth" "bare-auth-midline: should surface the auth token" + pass "fm-risk-tripwire matches a bare mid-line 'auth' but never 'authoritative'" +} + +test_adjacent_keywords_both_reported() { + # A boundary-consuming grep -o pattern drops the second word of an adjacent + # pair (the shared delimiter is eaten); whole-token matching reports both. + local case_dir out status + case_dir="$TMP_ROOT/adjacent-pair" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nRotate the session token on every login.\n\n# Setup\nx\n' \ + > "$case_dir/data/task-x1/brief.md" + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + expect_code 1 "$status" "adjacent-pair: adjacent risk words must trip" + assert_contains "$out" "session" "adjacent-pair: should surface session" + assert_contains "$out" "token" "adjacent-pair: should surface the adjacent token too" + pass "fm-risk-tripwire reports both words of an adjacent risk pair" +} + +test_multiword_phrase_keyword_trips() { + local case_dir out status + case_dir="$TMP_ROOT/phrase-keyword" + mkdir -p "$case_dir/state" "$case_dir/data/task-x1" + printf '# Task\nEnforce access control and handle data deletion on the endpoint.\n\n# Setup\nx\n' \ + > "$case_dir/data/task-x1/brief.md" + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + expect_code 1 "$status" "phrase-keyword: a multi-word risk phrase must still trip" + assert_contains "$out" "access control" "phrase-keyword: should surface the access control phrase" + pass "fm-risk-tripwire still trips on multi-word risk phrases" +} + test_clean_brief_and_diff_passes +test_bare_auth_matches_but_authoritative_does_not +test_adjacent_keywords_both_reported +test_multiword_phrase_keyword_trips test_brief_keyword_trips_wire test_diff_path_trips_wire test_brief_only_mode_before_worktree_exists diff --git a/tests/fm-tier-guard.test.sh b/tests/fm-tier-guard.test.sh index e2000432a..41650372f 100644 --- a/tests/fm-tier-guard.test.sh +++ b/tests/fm-tier-guard.test.sh @@ -176,9 +176,25 @@ test_missing_meta_errors() { status=$? set -e - expect_code 1 "$status" "missing-meta: no meta file must error, not silently pass" + expect_code 2 "$status" "missing-meta: no meta file must exit 2 (setup error), not 1 (an escalation)" assert_contains "$out" "no meta for task" "missing-meta: should report the missing meta file" - pass "fm-tier-guard errors when the task has no recorded meta" + pass "fm-tier-guard errors distinctly (exit 2) when the task has no recorded meta" +} + +test_usage_error_exit_code() { + local status + set +e + FM_ROOT_OVERRIDE="$ROOT" "$TIER_GUARD" >/dev/null 2>&1 + status=$? + set -e + expect_code 2 "$status" "usage-empty-id: a malformed invocation must exit 2, not 1 (the escalation code)" + + set +e + FM_ROOT_OVERRIDE="$ROOT" "$TIER_GUARD" one two >/dev/null 2>&1 + status=$? + set -e + expect_code 2 "$status" "usage-extra-args: extra args must exit 2, not 1 (the escalation code)" + pass "fm-tier-guard uses a distinct exit code for malformed invocations" } test_trivial_tier_within_envelope_passes @@ -188,5 +204,6 @@ test_trivial_tier_stale_age_escalates test_non_trivial_tier_small_diff_no_ceiling test_any_tier_past_heavy_scale_escalates test_missing_meta_errors +test_usage_error_exit_code echo "# all fm-tier-guard tests passed" From 1a39e58a7a6b334273ea26da2cf37bda11877743 Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 20:10:08 -0400 Subject: [PATCH 16/41] no-mistakes(review): set exec bit on new tests, fix ultracode marker doc --- tests/fm-risk-tripwire.test.sh | 0 tests/fm-tier-guard.test.sh | 0 tests/fm-ultracode-guard.test.sh | 0 3 files changed, 0 insertions(+), 0 deletions(-) mode change 100644 => 100755 tests/fm-risk-tripwire.test.sh mode change 100644 => 100755 tests/fm-tier-guard.test.sh mode change 100644 => 100755 tests/fm-ultracode-guard.test.sh diff --git a/tests/fm-risk-tripwire.test.sh b/tests/fm-risk-tripwire.test.sh old mode 100644 new mode 100755 diff --git a/tests/fm-tier-guard.test.sh b/tests/fm-tier-guard.test.sh old mode 100644 new mode 100755 diff --git a/tests/fm-ultracode-guard.test.sh b/tests/fm-ultracode-guard.test.sh old mode 100644 new mode 100755 From 3ea93ed6c48b57e35cbe42986b89cd24d8e74d9f Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 21:33:44 -0400 Subject: [PATCH 17/41] no-mistakes(test): fix jq/git-version and test-env portability in failing tests --- bin/fm-teardown.sh | 71 +++++++++++++++++++++------------------------- 1 file changed, 33 insertions(+), 38 deletions(-) diff --git a/bin/fm-teardown.sh b/bin/fm-teardown.sh index f8929d5fe..cb6b5898d 100755 --- a/bin/fm-teardown.sh +++ b/bin/fm-teardown.sh @@ -274,48 +274,43 @@ pr_is_merged() { unpushed_patches_are_in_pr_head "$head" } -# Does HEAD's content already match the given ref? Checks every path HEAD changed -# since its merge-base with ref: if ref already matches HEAD at each of those paths, -# HEAD introduces nothing ref does not already contain (e.g. its change landed via -# squash). This isolates branch-only changes, so unrelated commits ref gained past -# the merge-base do not count as "added" - equivalent to a 3-way merge's resulting -# tree matching ref's tree, but built from plain diff/merge-base so it works on git -# older than 2.38 (which is what `git merge-tree --write-tree` requires). Returns -# non-zero when inconclusive (ref missing, or a real difference). -content_matches_ref() { - local ref=$1 base changed_file - git -C "$WT" rev-parse --quiet --verify "$ref^{commit}" >/dev/null 2>&1 || return 1 - base=$(git -C "$WT" merge-base "$ref" HEAD 2>/dev/null) || return 1 - local -a changed=() - while IFS= read -r -d '' changed_file; do - changed+=("$changed_file") - done < <(git -C "$WT" diff --name-only -z "$base" HEAD -- 2>/dev/null) - [ "${#changed[@]}" -gt 0 ] || return 0 - git -C "$WT" diff --quiet "$ref" HEAD -- "${changed[@]}" 2>/dev/null -} - -# Is the branch's content already present in the up-to-date default branch? Checks -# two candidate refs, succeeding if either already contains HEAD's content: the -# fetched remote-tracking ref refs/remotes/origin/$name (the normal PR-merged case), -# and the LOCAL refs/heads/$name (a worktree and its project's primary checkout are -# linked worktrees of the same repository, so this local ref already reflects a -# firstmate-performed local fast-forward merge, even for a project whose delivery -# mode is not local-only - e.g. a PR closed unmerged upstream but the branch was -# landed into local main by hand on the captain's approval). A fetch failure only -# rules out the remote-tracking path; the local ref is still tried. Returns -# non-zero only when neither candidate ref exists or matches. +# Is the branch's content already present in the up-to-date default branch? Fetches +# first, then 3-way merges the default branch with HEAD: when HEAD introduces nothing +# the default branch does not already contain (e.g. its change landed via squash) the +# merged tree equals the default branch's tree. This isolates branch-only changes, so +# unrelated commits the default branch gained past the merge-base do not count as +# "added". Returns non-zero when inconclusive (no default ref, or a merge conflict), +# so the caller refuses rather than guesses. content_in_default() { - local name + local name ref default_tree base merged_tree tmpdir rc name=$(default_branch) || return 1 if git -C "$WT" remote get-url origin >/dev/null 2>&1; then - if git -C "$WT" fetch --quiet origin "+refs/heads/$name:refs/remotes/origin/$name" >/dev/null 2>&1; then - content_matches_ref "refs/remotes/origin/$name" && return 0 - fi - fi - if git -C "$WT" rev-parse --quiet --verify "refs/heads/$name" >/dev/null 2>&1; then - content_matches_ref "refs/heads/$name" && return 0 + git -C "$WT" fetch --quiet origin "+refs/heads/$name:refs/remotes/origin/$name" >/dev/null 2>&1 || return 1 + ref="refs/remotes/origin/$name" + elif git -C "$WT" rev-parse --quiet --verify "refs/heads/$name" >/dev/null 2>&1; then + ref="refs/heads/$name" + else + return 1 fi - return 1 + default_tree=$(git -C "$WT" rev-parse --quiet --verify "$ref^{tree}" 2>/dev/null) || return 1 + [ -n "$default_tree" ] || return 1 + base=$(git -C "$WT" merge-base "$ref" HEAD 2>/dev/null) || return 1 + # 3-way merge the default branch (ours) with HEAD (theirs) off their merge + # base into a throwaway index: when HEAD introduces nothing the default branch + # does not already contain, the merged tree equals the default branch's tree. + # read-tree + write-tree is used instead of `git merge-tree --write-tree` + # (git >= 2.38 only) so this stays portable to older git; a real merge + # conflict leaves unmerged index entries, write-tree fails, and the check + # stays inconclusive so the caller refuses rather than guesses. + tmpdir=$(mktemp -d 2>/dev/null) || return 1 + GIT_INDEX_FILE="$tmpdir/index" git -C "$WT" read-tree -im --aggressive \ + "$base^{tree}" "$ref^{tree}" "HEAD^{tree}" 2>/dev/null \ + && merged_tree=$(GIT_INDEX_FILE="$tmpdir/index" git -C "$WT" write-tree 2>/dev/null) + rc=$? + rm -rf "$tmpdir" + [ "$rc" -eq 0 ] || return 1 + [ -n "$merged_tree" ] || return 1 + [ "$merged_tree" = "$default_tree" ] } # Has the worktree's committed work actually LANDED, though its commits are not From 0e417d0d0ea50ccb8b35f60f7419ab040af2b405 Mon Sep 17 00:00:00 2001 From: Michael Date: Thu, 9 Jul 2026 22:02:59 -0400 Subject: [PATCH 18/41] no-mistakes(document): sync docs with new resource-tiering guardrail scripts --- docs/architecture.md | 7 +++++++ docs/configuration.md | 5 +++++ docs/scripts.md | 3 +++ 3 files changed, 15 insertions(+) diff --git a/docs/architecture.md b/docs/architecture.md index 4e71a6a41..fa385cbcd 100644 --- a/docs/architecture.md +++ b/docs/architecture.md @@ -130,6 +130,13 @@ Secondmate launches are exempt because they resolve the secondmate harness and a Unsupported effort values are still recorded in task meta when passed to `fm-spawn.sh`, but the launch template omits any effort flag that the selected harness does not accept. That keeps spawn launch compatible across claude, codex, grok, pi, and opencode while preserving the requested profile for later audit. +A crew-dispatch profile can also carry `ultracode: true` with an optional `ultracode_role` (today only `independent-review`): a task matched to that profile must get a genuinely independent second pass on its finished diff, dispatched as its own separate task rather than a sub-task the implementing crewmate spawns, before it can go PR-ready. +Three mechanical guardrails back this tiering, each a structurally separate check from the natural-language rule match so a misclassified task cannot slip a cheaper tier on one judgment call. +`bin/fm-risk-tripwire.sh` greps a task's brief text and, once code exists, its changed file paths for migration, auth, schema, and security signals; a hit floors the task to the safety-critical profile (`opus`/`xhigh`, ultracode `independent-review`) regardless of which rule was matched. +`bin/fm-tier-guard.sh` is a read-only Validate-time check that escalates a trivial (`haiku`/`low`) task in place once its actual diff size or elapsed time outgrows that tier's envelope, plus a general heavy-scale ceiling that escalates any tier once its diff crosses it; it reuses `bin/fm-review-diff.sh --stat` for the authoritative diff size. +`bin/fm-ultracode-guard.sh` tracks the ultracode requirement through a plain `state/.ultracode` marker (`role=`, then `reviewed_by=`), the same marker convention as `state/.afk`: it flags the requirement right after spawn (also on a risk-tripwire hit), records a distinct separately-dispatched reviewer task as the independent review, and refuses to let an ultracode-flagged task go PR-ready until that review is recorded. +None of the three modify `fm-spawn.sh` or task meta. + ## Optional secondmates `data/secondmates.md` records persistent domain supervisors with natural-language scopes, project clone lists, and home paths. diff --git a/docs/configuration.md b/docs/configuration.md index c9f459e8f..1fcf64efe 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -395,6 +395,11 @@ FM_CLASSIFY_PAUSED_VERB=paused # leading status verb for a declared external FM_STALE_ESCALATE_SECS=240 # idle seconds before a provably-working stale pane escalates; stale panes whose crew is not provably working surface immediately unless they declare the pause verb FM_PAUSE_RESURFACE_SECS=3600 # seconds before an idle declared external wait re-surfaces for a recheck in the watcher or away-mode daemon FM_WEDGE_DEMAND_INSPECT_COUNT=3 # consecutive provably-working stale escalations on the same unchanged pane before demand-deep-inspection is added +FM_TIER_TRIVIAL_MAX_LINES=30 # fm-tier-guard.sh: changed-line count a trivial (haiku/low) task may reach before escalation fires +FM_TIER_TRIVIAL_MAX_FILES=2 # fm-tier-guard.sh: changed-file count a trivial (haiku/low) task may reach before escalation fires +FM_TIER_TRIVIAL_MAX_AGE_SECONDS=1800 # fm-tier-guard.sh: elapsed seconds a trivial (haiku/low) task may run before escalation fires (30 min) +FM_TIER_HEAVY_MIN_LINES=400 # fm-tier-guard.sh: changed-line count that escalates any tier via the general heavy-scale ceiling +FM_TIER_HEAVY_MIN_FILES=8 # fm-tier-guard.sh: changed-file count that escalates any tier via the general heavy-scale ceiling FM_WATCH_TRIAGE_LOG_MAX_BYTES=262144 # size cap for the watcher's absorbed-wake debug log FM_FLEET_SYNC_BOOTSTRAP_TIMEOUT= # optional seconds allowed for bootstrap's best-effort clone refresh; unset/blank defaults to max(20, 5 + 3 * origin-backed-project-count) FM_FLEET_PRUNE=1 # set to 0 to skip pruning local branches whose upstream is gone diff --git a/docs/scripts.md b/docs/scripts.md index 1d8fc5810..71bfa3e1b 100644 --- a/docs/scripts.md +++ b/docs/scripts.md @@ -27,6 +27,9 @@ The shared no-mistakes gate refusal used by `fm-spawn.sh`, `fm-send.sh`, and `fm | `fm-home-seed.sh` | Transactionally provision a secondmate home and maintain `data/secondmates.md` | | `fm-spawn.sh` | Spawn crewmates, scouts, `id=repo` batches, and secondmates on the resolved harness and runtime backend; ship/scout spawns also derive and record a per-worktree docker-compose project name (`compose_project=`; docs/configuration.md "Docker Compose project isolation") | | `fm-dispatch-select.sh` | Resolve a matched crew-dispatch rule to one concrete profile, owning `quota-balanced` selection | +| `fm-tier-guard.sh` | Mechanically check whether a dispatched task's diff size or elapsed time has outgrown its assigned model/effort tier (guardrail #1) | +| `fm-risk-tripwire.sh` | Mechanically flag migration/auth/schema/security signals in a task's brief or changed files, overriding a misclassified cheap tier (guardrail #2) | +| `fm-ultracode-guard.sh` | Mechanically confirm an independent second-pass review ran on an ultracode-flagged task before PR-ready (guardrail #3) | | `fm-backend.sh` | Runtime-backend selection, meta helpers, selector resolution, and operation dispatch | | `fm-backend-hometag-lib.sh` | Shared per-installation home-tag derivation for zellij tab and cmux workspace titles | | `fm-composer-lib.sh` | Single fleet-wide owner of composer-content classification for all backends | From 1d48743c36b30705fc3bd0957c4a7a5044d3eeae Mon Sep 17 00:00:00 2001 From: Michael Date: Fri, 10 Jul 2026 13:24:55 -0400 Subject: [PATCH 19/41] no-mistakes(review): exclude herdr-lab brief block from risk-tripwire scan --- bin/fm-risk-tripwire.sh | 12 ++++++--- tests/fm-risk-tripwire.test.sh | 49 ++++++++++++++++++++++++++++++++-- 2 files changed, 56 insertions(+), 5 deletions(-) diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh index 3e8f1b585..9fc0058cc 100755 --- a/bin/fm-risk-tripwire.sh +++ b/bin/fm-risk-tripwire.sh @@ -83,8 +83,14 @@ PHRASE_REGEX='(access[[:space:]]+control|data[[:space:]]+deletion|bulk[[:space:] # next scaffold section heading), never the fixed scaffold boilerplate # bin/fm-brief.sh writes into every brief - that boilerplate contains benign # words like "authoritative" and "future session" that would otherwise trip the -# wire on every task. The section boundary is one of the scaffold's own known -# headings (# Setup/# Rules/# Project memory/# Definition of done), not any +# wire on every task. The Herdr block bin/fm-brief.sh injects immediately after +# # Task (either the # Herdr isolation ... hard-safety contract, whose text is +# dense with "session"/"--session", or the # Herdr lifecycle declaration ... +# not-enabled stub) is scaffold boilerplate too, so its heading is a boundary as +# well; without it every --herdr-lab brief would trip on the contract's own +# "session" wording rather than the real task text. The section boundary is one +# of the scaffold's own known headings (# Herdr isolation .../# Herdr lifecycle +# declaration .../# Setup/# Rules/# Project memory/# Definition of done), not any # column-0 "# " line: a shell/code comment embedded in the task body (e.g. # "# then run the schema migration") must NOT end the scan, or the risk words # after it are silently dropped - the dangerous direction for a safety floor. @@ -100,7 +106,7 @@ brief_task_body() { { blank = ($0 ~ /^[[:space:]]*$/) } /^```/ { fence = !fence; prevblank = blank; next } !fence && /^# Task[[:space:]]*$/ { intask=1; prevblank = blank; next } - intask && !fence && prevblank && /^# (Setup|Rules|Project memory|Definition of done)[[:space:]]*$/ { intask=0 } + intask && !fence && prevblank && /^# (Herdr isolation.*|Herdr lifecycle declaration.*|Setup|Rules|Project memory|Definition of done)[[:space:]]*$/ { intask=0 } intask { print } { prevblank = blank } ' "$1") diff --git a/tests/fm-risk-tripwire.test.sh b/tests/fm-risk-tripwire.test.sh index b29d0ac2a..4053ff14f 100755 --- a/tests/fm-risk-tripwire.test.sh +++ b/tests/fm-risk-tripwire.test.sh @@ -51,12 +51,15 @@ run_tripwire() { # Scaffold a real ship brief via bin/fm-brief.sh, then substitute the {TASK} # placeholder with the given task text - exercising the actual scaffold -# boilerplate rather than a hand-written stand-in. +# boilerplate rather than a hand-written stand-in. Any extra args (e.g. +# --herdr-lab) are forwarded to fm-brief.sh so tests cover the real injected +# sections, not a stand-in. scaffold_brief() { local case_dir=$1 task=$2 brief + shift 2 mkdir -p "$case_dir/state" FM_ROOT_OVERRIDE="$ROOT" FM_DATA_OVERRIDE="$case_dir/data" FM_STATE_OVERRIDE="$case_dir/state" \ - "$ROOT/bin/fm-brief.sh" task-x1 someproject >/dev/null 2>&1 + "$ROOT/bin/fm-brief.sh" task-x1 someproject "$@" >/dev/null 2>&1 brief="$case_dir/data/task-x1/brief.md" sed "s|{TASK}|$task|" "$brief" > "$brief.tmp" && mv "$brief.tmp" "$brief" } @@ -179,6 +182,46 @@ test_scaffolded_brief_risky_task_still_trips() { pass "fm-risk-tripwire still scans the task body of a scaffolded brief" } +test_herdr_lab_boilerplate_does_not_trip() { + # The --herdr-lab contract fm-brief.sh injects between # Task and # Setup is + # dense with "session"/"--session"; it is scaffold boilerplate, so a benign + # task must not trip on it. + local case_dir out status + case_dir="$TMP_ROOT/scaffold-herdr-clean" + scaffold_brief "$case_dir" "Fix a typo in the CLI help text." --herdr-lab + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1 2>&1) + status=$? + set -e + + expect_code 0 "$status" "scaffold-herdr-clean: the --herdr-lab contract's own 'session' text must not trip a benign task" + [ -z "$out" ] || fail "scaffold-herdr-clean: expected no RISK output from --herdr-lab boilerplate, got: $out" + pass "fm-risk-tripwire does not trip on --herdr-lab scaffold boilerplate" +} + +test_herdr_lab_risky_task_still_trips() { + # The Herdr block is now a scan boundary, so its "session" text is excluded - + # but the real task body between # Task and the Herdr heading must still be + # scanned, and the boilerplate's "session" must not leak into the hit list. + local case_dir out status + case_dir="$TMP_ROOT/scaffold-herdr-risky" + scaffold_brief "$case_dir" "Rotate the payment credentials and run the schema migration." --herdr-lab + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1) + status=$? + set -e + + expect_code 1 "$status" "scaffold-herdr-risky: a risk-worded task body must still trip under --herdr-lab" + assert_contains "$out" "RISK: brief for task-x1" "scaffold-herdr-risky: should name the brief hit" + assert_contains "$out" "payment" "scaffold-herdr-risky: should surface the real task-body term" + case "$out" in + *session*) fail "scaffold-herdr-risky: the Herdr boilerplate's 'session' must not leak into the hit list, got: $out" ;; + esac + pass "fm-risk-tripwire scans the task body but excludes the --herdr-lab Herdr block" +} + test_word_boundary_avoids_substring_false_positive() { local case_dir out status case_dir="$TMP_ROOT/word-boundary" @@ -581,6 +624,8 @@ test_brief_only_mode_before_worktree_exists test_nothing_to_check_errors test_scaffolded_brief_boilerplate_does_not_trip test_scaffolded_brief_risky_task_still_trips +test_herdr_lab_boilerplate_does_not_trip +test_herdr_lab_risky_task_still_trips test_word_boundary_avoids_substring_false_positive test_inflected_keyword_still_trips test_supervision_bin_path_does_not_trip From 4da224bb13b7d0d1d7d23459d9e49f93a3dfb1b2 Mon Sep 17 00:00:00 2001 From: Michael Date: Fri, 10 Jul 2026 13:36:11 -0400 Subject: [PATCH 20/41] no-mistakes(review): clean up orphan .ultracode state marker on teardown --- bin/fm-teardown.sh | 2 +- tests/fm-teardown.test.sh | 31 +++++++++++++++++++++++++++++++ 2 files changed, 32 insertions(+), 1 deletion(-) diff --git a/bin/fm-teardown.sh b/bin/fm-teardown.sh index cb6b5898d..9c412cac4 100755 --- a/bin/fm-teardown.sh +++ b/bin/fm-teardown.sh @@ -1057,7 +1057,7 @@ fm_backend_clear_transition "$BACKEND" "$STATE" "$T" || true # Remove the per-task temp root (/tmp/fm-/, incl. its gotmp/) recorded by spawn. # Read before the state-file rm below; empty (pre-fix tasks without tasktmp=) is a no-op. [ -n "$TASK_TMP" ] && rm -rf "$TASK_TMP" -rm -f "$STATE/$ID.status" "$STATE/$ID.turn-ended" "$STATE/$ID.check.sh" "$STATE/$ID.meta" "$STATE/$ID.pi-ext.ts" "$STATE/$ID.grok-turnend-token" +rm -f "$STATE/$ID.status" "$STATE/$ID.turn-ended" "$STATE/$ID.check.sh" "$STATE/$ID.meta" "$STATE/$ID.pi-ext.ts" "$STATE/$ID.grok-turnend-token" "$STATE/$ID.ultracode" if [ "$KIND" != scout ] && [ "$KIND" != secondmate ] && [ "$MODE" != local-only ]; then "$FM_ROOT/bin/fm-fleet-sync.sh" "$PROJ" || true fi diff --git a/tests/fm-teardown.test.sh b/tests/fm-teardown.test.sh index c1cc2b35a..bf9dbd7fc 100755 --- a/tests/fm-teardown.test.sh +++ b/tests/fm-teardown.test.sh @@ -642,6 +642,36 @@ test_no_mistakes_origin_remote_allows() { pass "no-mistakes worktree with HEAD on origin is torn down (no regression)" } +test_teardown_removes_per_task_state_markers() { + local case_dir rc + case_dir=$(make_case state-markers) + write_meta "$case_dir" no-mistakes ship + wt_commit "$case_dir" "shippable work" + git -C "$case_dir/wt" push -q origin fm/task-x1 + git -C "$case_dir/project" fetch -q origin + # Every per-task marker a teardown is responsible for reaping, including the + # ultracode flag written by fm-ultracode-guard.sh. + printf '%s\n' 'working: x' > "$case_dir/state/task-x1.status" + touch "$case_dir/state/task-x1.turn-ended" + printf '%s\n' 'exit 0' > "$case_dir/state/task-x1.check.sh" + printf '%s\n' 'ts' > "$case_dir/state/task-x1.pi-ext.ts" + printf '%s\n' 'tok' > "$case_dir/state/task-x1.grok-turnend-token" + printf '%s\n' 'role=independent-review' > "$case_dir/state/task-x1.ultracode" + + set +e + run_teardown "$case_dir" > "$case_dir/stdout" 2> "$case_dir/stderr" + rc=$? + set -e + + expect_code 0 "$rc" "state-markers: teardown should succeed when HEAD is on origin" + local marker + for marker in status turn-ended check.sh meta pi-ext.ts grok-turnend-token ultracode; do + [ ! -e "$case_dir/state/task-x1.$marker" ] \ + || fail "state-markers: teardown left orphan state/task-x1.$marker" + done + pass "teardown removes every per-task state marker including .ultracode (no leak)" +} + test_no_mistakes_truly_unpushed_refuses() { local case_dir rc case_dir=$(make_case nm-unpushed) @@ -1333,6 +1363,7 @@ test_teardown_manual_backend_prompts_hand_edit_even_when_tasks_axi_present test_local_only_truly_unpushed_refuses test_local_only_merged_to_local_main_allows test_no_mistakes_origin_remote_allows +test_teardown_removes_per_task_state_markers test_no_mistakes_truly_unpushed_refuses test_no_mistakes_merged_to_local_main_allows test_local_only_force_overrides_unpushed From 6e8e8bf3c77ff8561def0ac08aafd7355070175e Mon Sep 17 00:00:00 2001 From: Michael Date: Fri, 10 Jul 2026 14:10:07 -0400 Subject: [PATCH 21/41] no-mistakes(review): stop tiering guardrails passing silently when diff uncheckable --- bin/fm-risk-tripwire.sh | 25 +++++++++++++++-- bin/fm-tier-guard.sh | 20 ++++++++++++- tests/fm-risk-tripwire.test.sh | 51 ++++++++++++++++++++++++++++++++++ tests/fm-tier-guard.test.sh | 35 +++++++++++++++++++++++ 4 files changed, 128 insertions(+), 3 deletions(-) diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh index 9fc0058cc..aee4c71f6 100755 --- a/bin/fm-risk-tripwire.sh +++ b/bin/fm-risk-tripwire.sh @@ -17,8 +17,10 @@ # Exit codes: # 0 no risk signal found # 1 a RISK hit fired (one "RISK: " line per hit is printed) -# 2 could not check: a malformed invocation, or neither a brief nor a usable -# worktree/project exists yet (nothing to check) +# 2 could not check: a malformed invocation, neither a brief nor a usable +# worktree/project exists yet (nothing to check), or the worktree/project +# exist but their diff base could not be resolved, so the binding diff +# checkpoint could not run (warned to stderr, never a silent clean pass) # Distinct codes matter so a caller branching on $? cannot mistake a malformed # invocation for a real risk hit. A hit means: floor this task's model/effort to # the safety-critical profile (opus/xhigh, ultracode independent-review) @@ -162,6 +164,7 @@ path_is_risky() { FOUND=0 CHECKED=0 +DIFF_UNRESOLVED=0 BRIEF="$DATA/$ID/brief.md" if [ -f "$BRIEF" ]; then @@ -185,6 +188,7 @@ if [ -f "$META" ]; then PROJ=$(grep '^project=' "$META" | tail -1 | cut -d= -f2- || true) if [ -n "$WT" ] && [ -n "$PROJ" ] && [ -d "$WT" ] && [ -d "$PROJ" ]; then CHECKED=1 + diff_base_resolved=0 # The base resolution below duplicates bin/fm-review-diff.sh's origin fetch # and origin/ verification on purpose: this scan needs the changed # PATH LIST, but fm-review-diff.sh only exposes --stat (a diffstat), not a @@ -199,6 +203,7 @@ if [ -f "$META" ]; then git -C "$WT" rev-parse --verify --quiet "refs/remotes/origin/$DEFAULT^{commit}" >/dev/null 2>&1 && BASE="origin/$DEFAULT" fi if git -C "$WT" rev-parse --verify --quiet "$BASE^{commit}" >/dev/null 2>&1; then + diff_base_resolved=1 diff_paths=$(git -C "$WT" diff --name-only "$BASE...HEAD" -- 2>/dev/null || true) risky_paths= if [ -n "$diff_paths" ]; then @@ -216,6 +221,16 @@ if [ -f "$META" ]; then fi fi fi + # The worktree/project exist, so the binding diff checkpoint was expected to + # run - but the diff base could not be resolved (no default branch, or the + # base commit does not verify in the worktree). Do not let that read as a + # clean pass: warn loudly, like the sibling bin/fm-review-diff.sh, and mark + # the run not-checkable so a caller branching on $? floors the task rather + # than trusting a silent 0 - the safe direction for this guardrail. + if [ "$diff_base_resolved" -eq 0 ]; then + echo "warning: could not resolve a diff base for $ID (project $PROJ, default branch '${DEFAULT:-unresolved}'); the diff checkpoint did not run - reporting not-checkable (exit 2), not a clean pass" >&2 + DIFF_UNRESOLVED=1 + fi fi fi @@ -224,4 +239,10 @@ if [ "$CHECKED" -eq 0 ]; then exit 2 fi +# A real risk hit (exit 1) always wins over an unresolvable diff base; only when +# nothing tripped does an unresolvable base downgrade the run to could-not-check. +if [ "$FOUND" -eq 0 ] && [ "$DIFF_UNRESOLVED" -eq 1 ]; then + exit 2 +fi + exit "$FOUND" diff --git a/bin/fm-tier-guard.sh b/bin/fm-tier-guard.sh index 38234c4b0..db71e381c 100755 --- a/bin/fm-tier-guard.sh +++ b/bin/fm-tier-guard.sh @@ -87,10 +87,24 @@ is_trivial_tier() { return 1 } -STAT_OUT=$("$SCRIPT_DIR/fm-review-diff.sh" "$ID" --stat 2>/dev/null) || { +STAT_ERR=$(mktemp) +STAT_OUT=$("$SCRIPT_DIR/fm-review-diff.sh" "$ID" --stat 2>"$STAT_ERR") || { echo "error: fm-review-diff.sh failed for $ID" >&2 + rm -f "$STAT_ERR" exit 2 } +# fm-review-diff.sh still exits 0 but warns when it cannot resolve an open PR's +# head and falls back to the possibly-stale local branch, so the size measured +# below may under-count the real PR diff. Blanket-discarding that warning (the +# 2>/dev/null this replaces) made a degraded read indistinguishable from an +# authoritative one - the dangerous under-report direction for this guardrail. +# Detect only that specific marker (not all of fm-review-diff.sh's stderr, which +# also carries fm-guard.sh's supervision banner) and surface it below. +DIFF_MAY_LAG_PR=0 +if grep -q 'PR head unavailable' "$STAT_ERR"; then + DIFF_MAY_LAG_PR=1 +fi +rm -f "$STAT_ERR" FILES=0 LINES=0 @@ -127,4 +141,8 @@ if [ "$FILES" -gt "$HEAVY_MIN_FILES" ] || [ "$LINES" -gt "$HEAVY_MIN_LINES" ]; t FOUND=1 fi +if [ "$DIFF_MAY_LAG_PR" -eq 1 ]; then + echo "warning: the diff for $ID was sized against a possibly-stale local branch because the open PR head could not be resolved; the ceilings above may under-count the real PR diff - re-check against the current PR before trusting a within-envelope result" >&2 +fi + exit "$FOUND" diff --git a/tests/fm-risk-tripwire.test.sh b/tests/fm-risk-tripwire.test.sh index 4053ff14f..bccf020fa 100755 --- a/tests/fm-risk-tripwire.test.sh +++ b/tests/fm-risk-tripwire.test.sh @@ -599,6 +599,55 @@ test_adjacent_keywords_both_reported() { pass "fm-risk-tripwire reports both words of an adjacent risk pair" } +test_unresolvable_diff_base_is_not_a_clean_pass() { + # A meta with a real worktree/project but an unresolvable diff base (no default + # branch, no origin) must NOT silently read as a clean pass (exit 0). The + # binding second checkpoint could not run, so it must warn and report + # could-not-check (2), matching the sibling fm-tier-guard.sh/fm-review-diff.sh. + local case_dir out status + case_dir="$TMP_ROOT/unresolvable-base" + mkdir -p "$case_dir/state" "$case_dir/wt" + git init -q "$case_dir/project" + + fm_write_meta "$case_dir/state/task-x1.meta" \ + "window=fm-task-x1" \ + "worktree=$case_dir/wt" \ + "project=$case_dir/project" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1 2>&1) + status=$? + set -e + + expect_code 2 "$status" "unresolvable-base: an unresolvable diff base must read as could-not-check (2), not a clean pass (0)" + assert_contains "$out" "could not resolve a diff base" "unresolvable-base: should warn that the diff checkpoint did not run" + pass "fm-risk-tripwire reports could-not-check when the diff base is unresolvable, not a clean pass" +} + +test_unresolvable_diff_base_still_reports_brief_hit() { + # A brief risk hit must still win (exit 1) even when the diff base is + # unresolvable - the risk floor beats the could-not-check downgrade. + local case_dir out status + case_dir="$TMP_ROOT/unresolvable-base-brief-hit" + mkdir -p "$case_dir/state" "$case_dir/wt" "$case_dir/data/task-x1" + git init -q "$case_dir/project" + printf 'Add a data migration for the new billing schema.\n' > "$case_dir/data/task-x1/brief.md" + + fm_write_meta "$case_dir/state/task-x1.meta" \ + "window=fm-task-x1" \ + "worktree=$case_dir/wt" \ + "project=$case_dir/project" + + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" FM_DATA_OVERRIDE="$case_dir/data" "$TRIPWIRE" task-x1 2>/dev/null) + status=$? + set -e + + expect_code 1 "$status" "unresolvable-base-brief-hit: a brief risk hit must still win (1) even when the diff base is unresolvable" + assert_contains "$out" "RISK: brief for task-x1" "unresolvable-base-brief-hit: should still name the brief hit" + pass "fm-risk-tripwire still reports a brief hit (1) when the diff base is unresolvable" +} + test_multiword_phrase_keyword_trips() { local case_dir out status case_dir="$TMP_ROOT/phrase-keyword" @@ -617,6 +666,8 @@ test_multiword_phrase_keyword_trips() { test_clean_brief_and_diff_passes test_bare_auth_matches_but_authoritative_does_not test_adjacent_keywords_both_reported +test_unresolvable_diff_base_is_not_a_clean_pass +test_unresolvable_diff_base_still_reports_brief_hit test_multiword_phrase_keyword_trips test_brief_keyword_trips_wire test_diff_path_trips_wire diff --git a/tests/fm-tier-guard.test.sh b/tests/fm-tier-guard.test.sh index 41650372f..0df6a3b0b 100755 --- a/tests/fm-tier-guard.test.sh +++ b/tests/fm-tier-guard.test.sh @@ -167,6 +167,40 @@ test_any_tier_past_heavy_scale_escalates() { pass "fm-tier-guard escalates any tier once its diff crosses the general heavy-scale ceiling" } +test_unresolvable_pr_head_surfaces_degraded_diff_warning() { + # When state records pr= but the PR head cannot be resolved, fm-review-diff.sh + # falls back to the possibly-stale local branch and warns (still exit 0). The + # guard must NOT swallow that: the size it measured may under-count the real PR + # diff, so the degradation has to stay visible instead of reading as a clean, + # authoritative within-envelope pass - the same silent-reads-clean class as the + # sibling fm-risk-tripwire.sh unresolvable-base guard. + local case_dir out err status + case_dir=$(make_case degraded-diff) + printf 'base\nsmall change\n' > "$case_dir/wt/feature.txt" + git -C "$case_dir/wt" add feature.txt + git -C "$case_dir/wt" commit -qm "small local edit" + # pr= recorded, but pr_head is a bogus sha and origin has no matching PR ref, + # so resolve_pr_head fails and fm-review-diff.sh takes the stale-branch path. + write_task_meta "$case_dir" task-x1 "model=claude-haiku-4-5" "effort=low" \ + "pr=https://github.com/o/r/pull/7" \ + "pr_head=0000000000000000000000000000000000000000" + + local errfile; errfile="$case_dir/stderr.txt" + set +e + out=$(FM_ROOT_OVERRIDE="$ROOT" FM_STATE_OVERRIDE="$case_dir/state" "$TIER_GUARD" task-x1 2>"$errfile") + status=$? + set -e + err=$(cat "$errfile") + + expect_code 0 "$status" "degraded-diff: a small stale-branch diff still measures within envelope (exit 0)" + [ -z "$out" ] || fail "degraded-diff: expected no ESCALATE on stdout, got: $out" + assert_contains "$err" "may under-count the real PR diff" "degraded-diff: the degraded measurement must be surfaced, not silently swallowed" + case "$err" in + *"WATCHER DOWN"*) fail "degraded-diff: must not leak fm-guard.sh's supervision banner from fm-review-diff.sh stderr, got: $err" ;; + esac + pass "fm-tier-guard surfaces a degraded (PR-head-unavailable) diff measurement instead of swallowing it" +} + test_missing_meta_errors() { local case_dir out status case_dir=$(make_case missing-meta) @@ -203,6 +237,7 @@ test_trivial_tier_exceeds_file_ceiling_escalates test_trivial_tier_stale_age_escalates test_non_trivial_tier_small_diff_no_ceiling test_any_tier_past_heavy_scale_escalates +test_unresolvable_pr_head_surfaces_degraded_diff_warning test_missing_meta_errors test_usage_error_exit_code From 60933a4e0489f49ad2ee895552d634efad1ee7e9 Mon Sep 17 00:00:00 2001 From: Michael Date: Fri, 10 Jul 2026 14:49:03 -0400 Subject: [PATCH 22/41] no-mistakes(review): pass ultracode through dispatch, tighten tiering guardrail coverage --- bin/fm-dispatch-select.sh | 8 ++++++-- bin/fm-risk-tripwire.sh | 2 +- bin/fm-tier-guard.sh | 6 +++++- 3 files changed, 12 insertions(+), 4 deletions(-) diff --git a/bin/fm-dispatch-select.sh b/bin/fm-dispatch-select.sh index c626e9a95..dfe100379 100755 --- a/bin/fm-dispatch-select.sh +++ b/bin/fm-dispatch-select.sh @@ -115,7 +115,9 @@ first_profile() { def clean($p): {harness: $p.harness} + (if ($p.model? | type) == "string" then {model: $p.model} else {} end) - + (if ($p.effort? | type) == "string" then {effort: $p.effort} else {} end); + + (if ($p.effort? | type) == "string" then {effort: $p.effort} else {} end) + + (if ($p.ultracode? | type) == "boolean" then {ultracode: $p.ultracode} else {} end) + + (if ($p.ultracode_role? | type) == "string" then {ultracode_role: $p.ultracode_role} else {} end); clean(.[0]) ' } @@ -169,7 +171,9 @@ selection=$(printf '%s\n' "$quota_json" | jq -ec \ def clean($p): {harness: $p.harness} + (if ($p.model? | type) == "string" then {model: $p.model} else {} end) - + (if ($p.effort? | type) == "string" then {effort: $p.effort} else {} end); + + (if ($p.effort? | type) == "string" then {effort: $p.effort} else {} end) + + (if ($p.ultracode? | type) == "boolean" then {ultracode: $p.ultracode} else {} end) + + (if ($p.ultracode_role? | type) == "string" then {ultracode_role: $p.ultracode_role} else {} end); def provider_for($h): [.providers[]? | select(.provider == $h)][0]; def general_ids($h): if $h == "claude" then ["five_hour", "seven_day"] diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh index aee4c71f6..e684eff57 100755 --- a/bin/fm-risk-tripwire.sh +++ b/bin/fm-risk-tripwire.sh @@ -138,7 +138,7 @@ brief_task_body() { # NO delimiter (e.g. "authsetup.rb") is not matched, because catching it would # require substring matching again and reopen exactly those false positives; the # brief-text scan and the delimiter tokenization below cover the realistic cases. -PATH_STRONG_REGEX='^(auth|authn|authz|authoriz(e|ed|es|ing|ation|ations|er|ers)?|authenticat(e|ed|es|ing|ion|ions|or|ors)?|migrat(e|ed|es|ing|ion|ions)|schema|schemas|secret|secrets|credential|credentials|payment|payments|billing|security)$' +PATH_STRONG_REGEX='^(auth|authn|authz|authoriz(e|ed|es|ing|ation|ations|er|ers)?|authenticat(e|ed|es|ing|ion|ions|or|ors)?|migrat(e|ed|es|ing|ion|ions)|schema|schemas|secret|secrets|credential|credentials|payment|payments|billing|security|password|passwords|token|tokens|permission|permissions|encrypt(s|ed|ing|ion|ions)?|decrypt(s|ed|ing|ion|ions)?)$' PATH_WEAK_REGEX='^(session|sessions)$' path_is_risky() { local path=$1 comp base diff --git a/bin/fm-tier-guard.sh b/bin/fm-tier-guard.sh index db71e381c..934a5bfb4 100755 --- a/bin/fm-tier-guard.sh +++ b/bin/fm-tier-guard.sh @@ -82,7 +82,11 @@ tier_guard_stat_mtime() { is_trivial_tier() { case "$MODEL" in - *haiku*) [ "$EFFORT" = low ] && return 0 ;; + *haiku*) + if [ "$EFFORT" = low ] || [ "$EFFORT" = default ]; then + return 0 + fi + ;; esac return 1 } From bdea9a8f4751a0617e20ed7166d3cb021c2e33da Mon Sep 17 00:00:00 2001 From: Michael Date: Fri, 10 Jul 2026 15:41:28 -0400 Subject: [PATCH 23/41] no-mistakes(test): skip handoff tests on incompatible tasks-axi; fix herdr server-race --- tests/fm-backend-herdr.test.sh | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-) diff --git a/tests/fm-backend-herdr.test.sh b/tests/fm-backend-herdr.test.sh index f3100d484..77484a693 100755 --- a/tests/fm-backend-herdr.test.sh +++ b/tests/fm-backend-herdr.test.sh @@ -39,6 +39,14 @@ COUNT_FILE="$RESP/.count" for a in "$@"; do printf '\x1f%s' "$a"; done printf '\n' } >> "$LOG" +# `herdr server` is a fire-and-forget launch: server_ensure backgrounds it and +# discards its output, so its timing relative to the immediately-following status +# poll is non-deterministic. It must NOT consume a canned-response slot, or a lost +# race with that poll would steal a status response and desync the whole sequence. +# It is still logged above, so the "did it start the server" assertion holds. +if [ "${1:-}" = server ]; then + exit 0 +fi if [ "${1:-}" = status ] && [ "${2:-}" = --json ] && [ "${FM_HERDR_SCRIPT_STATUS:-0}" != 1 ]; then printf '{"client":{"version":"0.7.1","protocol":14},"server":{"running":true}}\n' exit 0 @@ -300,14 +308,15 @@ test_container_ensure_starts_server_and_workspace() { printf '{"client":{"version":"0.7.1","protocol":14}}\n' > "$resp/1.out" # 2: server_ensure's status --json check -> not running printf '{"server":{"running":false}}\n' > "$resp/2.out" - # 3: `herdr server` backgrounded launch - no meaningful output - # 4: server_ensure poll -> now running - printf '{"server":{"running":true}}\n' > "$resp/4.out" - # 5: workspace list -> empty (no "firstmate" workspace yet) - printf '{"result":{"workspaces":[]}}\n' > "$resp/5.out" - # 6: workspace create -> w1, seeding default tab w1:t9 (real herdr returns + # `herdr server` is backgrounded and consumes no response slot (see the fake), + # so the very next scripted response is the server_ensure poll. + # 3: server_ensure poll -> now running + printf '{"server":{"running":true}}\n' > "$resp/3.out" + # 4: workspace list -> empty (no "firstmate" workspace yet) + printf '{"result":{"workspaces":[]}}\n' > "$resp/4.out" + # 5: workspace create -> w1, seeding default tab w1:t9 (real herdr returns # the seeded tab/pane ids in the SAME response - verified empirically). - printf '{"result":{"workspace":{"workspace_id":"w1","label":"firstmate"},"tab":{"tab_id":"w1:t9"},"root_pane":{"pane_id":"w1:p9"}}}\n' > "$resp/6.out" + printf '{"result":{"workspace":{"workspace_id":"w1","label":"firstmate"},"tab":{"tab_id":"w1:t9"},"root_pane":{"pane_id":"w1:p9"}}}\n' > "$resp/5.out" fb=$(make_herdr_fakebin "$dir") out=$( PATH="$fb:$PATH" FM_HERDR_LOG="$log" FM_HERDR_RESPONSES="$resp" FM_HERDR_SCRIPT_STATUS=1 HERDR_SESSION=fmtest \ bash -c '. "$0/bin/backends/herdr.sh"; fm_backend_herdr_container_ensure /tmp' "$ROOT" ) From 805eab12940b81ac02a8e7021c9890ef7ec9770e Mon Sep 17 00:00:00 2001 From: Michael Date: Fri, 10 Jul 2026 17:15:34 -0400 Subject: [PATCH 24/41] no-mistakes: apply CI fixes --- bin/fm-ultracode-guard.sh | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/bin/fm-ultracode-guard.sh b/bin/fm-ultracode-guard.sh index ed0de277a..485510fa3 100755 --- a/bin/fm-ultracode-guard.sh +++ b/bin/fm-ultracode-guard.sh @@ -45,7 +45,10 @@ fi CMD=${1:-} ID=${2:-} -[ -n "$CMD" ] && [ -n "$ID" ] || { usage; exit 1; } +if [ -z "$CMD" ] || [ -z "$ID" ]; then + usage + exit 1 +fi MARKER="$STATE/$ID.ultracode" From 5459c597ed5b9c1e09734e1db30be8a7e18f76f6 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 15:24:41 -0400 Subject: [PATCH 25/41] fix(teardown): check both origin and local ref for landed content The read-tree portability rewrite restructured content_in_default() into an if/elif that checks exactly one candidate ref: the fetched origin ref when a remote exists, falling back to the local ref only when there is no remote at all. That silently dropped the case a no-mistakes-mode project with a closed, unmerged PR but work fast-forwarded into local main by hand - the fetched origin ref exists but does not contain the work, so the local ref is never even checked. Restore checking both candidates in sequence via a shared content_matches_ref helper, matching the pre-rewrite behavior tests/fm-teardown.test.sh's local-land regression test expects. --- bin/fm-teardown.sh | 58 ++++++++++++++++++++++++++-------------------- 1 file changed, 33 insertions(+), 25 deletions(-) diff --git a/bin/fm-teardown.sh b/bin/fm-teardown.sh index 9c412cac4..b59dd848d 100755 --- a/bin/fm-teardown.sh +++ b/bin/fm-teardown.sh @@ -274,34 +274,18 @@ pr_is_merged() { unpushed_patches_are_in_pr_head "$head" } -# Is the branch's content already present in the up-to-date default branch? Fetches -# first, then 3-way merges the default branch with HEAD: when HEAD introduces nothing -# the default branch does not already contain (e.g. its change landed via squash) the -# merged tree equals the default branch's tree. This isolates branch-only changes, so -# unrelated commits the default branch gained past the merge-base do not count as -# "added". Returns non-zero when inconclusive (no default ref, or a merge conflict), -# so the caller refuses rather than guesses. -content_in_default() { - local name ref default_tree base merged_tree tmpdir rc - name=$(default_branch) || return 1 - if git -C "$WT" remote get-url origin >/dev/null 2>&1; then - git -C "$WT" fetch --quiet origin "+refs/heads/$name:refs/remotes/origin/$name" >/dev/null 2>&1 || return 1 - ref="refs/remotes/origin/$name" - elif git -C "$WT" rev-parse --quiet --verify "refs/heads/$name" >/dev/null 2>&1; then - ref="refs/heads/$name" - else - return 1 - fi +# Does the given ref's tree already contain everything HEAD introduces? 3-way merges +# the ref (ours) with HEAD (theirs) off their merge base into a throwaway index: when +# HEAD introduces nothing the ref does not already contain (e.g. its change landed via +# squash), the merged tree equals the ref's tree. read-tree + write-tree is used +# instead of `git merge-tree --write-tree` (git >= 2.38 only) so this stays portable to +# older git; a real merge conflict leaves unmerged index entries, write-tree fails, and +# the check stays inconclusive so the caller refuses rather than guesses. +content_matches_ref() { + local ref=$1 default_tree base merged_tree tmpdir rc default_tree=$(git -C "$WT" rev-parse --quiet --verify "$ref^{tree}" 2>/dev/null) || return 1 [ -n "$default_tree" ] || return 1 base=$(git -C "$WT" merge-base "$ref" HEAD 2>/dev/null) || return 1 - # 3-way merge the default branch (ours) with HEAD (theirs) off their merge - # base into a throwaway index: when HEAD introduces nothing the default branch - # does not already contain, the merged tree equals the default branch's tree. - # read-tree + write-tree is used instead of `git merge-tree --write-tree` - # (git >= 2.38 only) so this stays portable to older git; a real merge - # conflict leaves unmerged index entries, write-tree fails, and the check - # stays inconclusive so the caller refuses rather than guesses. tmpdir=$(mktemp -d 2>/dev/null) || return 1 GIT_INDEX_FILE="$tmpdir/index" git -C "$WT" read-tree -im --aggressive \ "$base^{tree}" "$ref^{tree}" "HEAD^{tree}" 2>/dev/null \ @@ -313,6 +297,30 @@ content_in_default() { [ "$merged_tree" = "$default_tree" ] } +# Is the branch's content already present in the up-to-date default branch? Checks +# two candidate refs, succeeding if either already contains HEAD's content: the +# fetched remote-tracking ref refs/remotes/origin/$name (the normal PR-merged case), +# and the LOCAL refs/heads/$name (a worktree and its project's primary checkout are +# linked worktrees of the same repository, so this local ref already reflects a +# firstmate-performed local fast-forward merge, even for a project whose delivery +# mode is not local-only - e.g. a PR closed unmerged upstream but the branch was +# landed into local main by hand on the captain's approval). A fetch failure only +# rules out the remote-tracking path; the local ref is still tried. Returns +# non-zero only when neither candidate ref exists or matches. +content_in_default() { + local name + name=$(default_branch) || return 1 + if git -C "$WT" remote get-url origin >/dev/null 2>&1; then + if git -C "$WT" fetch --quiet origin "+refs/heads/$name:refs/remotes/origin/$name" >/dev/null 2>&1; then + content_matches_ref "refs/remotes/origin/$name" && return 0 + fi + fi + if git -C "$WT" rev-parse --quiet --verify "refs/heads/$name" >/dev/null 2>&1; then + content_matches_ref "refs/heads/$name" && return 0 + fi + return 1 +} + # Has the worktree's committed work actually LANDED, though its commits are not # reachable from any remote-tracking branch? True when a merged PR proves the # current local work is contained in the PR head, OR the content is already in the From 6c5fd06d6bb87d049d2f3453bb246e4c5a1fb59c Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 20:12:06 -0400 Subject: [PATCH 26/41] feat: auto-sync live NAS deployments after a landed ship task A merged PR does not, by itself, reach the live site: deployed apps run from a separate checkout under /mnt/nas/experiments//, managed by pm2, independent from the projects/ clone used for development. This gap required a manual pull+restart to notice and fix. bin/fm-nas-deploy-sync.sh fast-forwards a project's live NAS checkout and restarts its pm2 process(es), looked up from a captain-private data/nas-deployments.md mapping (best-effort, non-fatal, mirrors fm-fleet-sync.sh's dirty/diverged safety exactly). bin/fm-teardown.sh now calls it automatically after a landed ship task, including local-only merges. --- bin/fm-nas-deploy-sync.sh | 191 ++++++++++++++++++ bin/fm-teardown.sh | 9 + tests/fm-nas-deploy-sync.test.sh | 335 +++++++++++++++++++++++++++++++ 3 files changed, 535 insertions(+) create mode 100755 bin/fm-nas-deploy-sync.sh create mode 100755 tests/fm-nas-deploy-sync.test.sh diff --git a/bin/fm-nas-deploy-sync.sh b/bin/fm-nas-deploy-sync.sh new file mode 100755 index 000000000..8899703d0 --- /dev/null +++ b/bin/fm-nas-deploy-sync.sh @@ -0,0 +1,191 @@ +#!/usr/bin/env bash +# Best-effort, non-fatal deploy sync for a single project's live NAS checkout: a +# merged PR lands in the project's GitHub default branch, but the live app on this +# host runs from its own separate checkout under /mnt/nas/experiments//, +# managed by pm2 - that checkout does not update itself (see the 2026-07-04 entry +# in data/learnings.md). This script closes that gap after a landed ship task. +# +# Looks the project up in data/nas-deployments.md, a pipe-table with one row per +# deployed project: "| | | |" (comma- +# separated when more than one process serves the app). A project absent from +# that table - or a fresh checkout with no data/nas-deployments.md at all - has no +# known live deployment, which is not an error: prints one skip line and exits 0. +# +# When present: fetches the NAS checkout's origin and fast-forward-only merges to +# origin/, mirroring bin/fm-fleet-sync.sh's dirty/diverged safety exactly +# - never force, never discard local changes, never touch a checkout that is off +# its default branch, dirty, or diverged (reported as STUCK, left untouched). +# Restarts the recorded pm2 process(es) only when the merge actually advanced +# HEAD, then verifies each is back online via `pm2 jlist`. Always prints exactly +# one clear result line and exits 0 - this script is best-effort and must never +# fail or block its caller (see the post-teardown call in bin/fm-teardown.sh). +# +# Usage: fm-nas-deploy-sync.sh +# +# Overrides (test injection): FM_DATA_OVERRIDE points at an alternate data/ dir +# (same knob bin/fm-project-mode.sh uses); FM_NAS_DEPLOYMENTS_OVERRIDE points at +# an alternate mapping file directly. pm2 is invoked via PATH, so a fakebin shim +# ahead of it on PATH intercepts restart/list during tests - real tests must never +# reach the real /mnt/nas/experiments or a real pm2 process. +set -eu + +SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" +FM_ROOT="${FM_ROOT_OVERRIDE:-$(cd "$SCRIPT_DIR/.." && pwd)}" +FM_HOME="${FM_HOME:-${FM_ROOT_OVERRIDE:-$FM_ROOT}}" +DATA="${FM_DATA_OVERRIDE:-$FM_HOME/data}" +MAP="${FM_NAS_DEPLOYMENTS_OVERRIDE:-$DATA/nas-deployments.md}" + +usage() { + echo "usage: fm-nas-deploy-sync.sh " >&2 +} + +if [ "${1:-}" = "--help" ] || [ "${1:-}" = "-h" ]; then + usage + exit 0 +fi +[ $# -eq 1 ] || { usage; exit 1; } +NAME=$1 + +first_line() { + printf '%s\n' "$1" | sed -n '1s/[[:space:]]\{1,\}/ /g;1p' +} + +# lookup_row: print "|" for $NAME's row in $MAP, or +# nothing when the file is absent or has no matching row. +lookup_row() { + [ -f "$MAP" ] || return 0 + # $3 ~ /^\// excludes the header ("nas_repo_path") and separator ("---") rows - + # a real NAS path is always absolute - so a project literally named "project" + # cannot collide with the header's own column label. + awk -F'|' -v n="$NAME" ' + { for (i = 1; i <= NF; i++) { gsub(/^[ \t]+|[ \t]+$/, "", $i) } } + NF >= 4 && $2 == n && $3 ~ /^\// { print $3 "|" $4; exit } + ' "$MAP" +} + +row=$(lookup_row) +if [ -z "$row" ]; then + echo "$NAME: skipped: no recorded NAS deployment" + exit 0 +fi +NAS_PATH=${row%%|*} +PROCS=${row#*|} + +if [ ! -d "$NAS_PATH" ]; then + echo "$NAME: skipped: NAS path $NAS_PATH not a directory" + exit 0 +fi +if ! git -C "$NAS_PATH" rev-parse --is-inside-work-tree >/dev/null 2>&1; then + echo "$NAME: skipped: $NAS_PATH not a git repo" + exit 0 +fi +if ! git -C "$NAS_PATH" remote get-url origin >/dev/null 2>&1; then + echo "$NAME: skipped: no origin remote at $NAS_PATH" + exit 0 +fi + +fetch_output="" +if ! fetch_output=$(git -C "$NAS_PATH" fetch origin --quiet 2>&1); then + reason="fetch failed" + [ -z "$fetch_output" ] || reason="$reason: $(first_line "$fetch_output")" + echo "$NAME: skipped: $reason" + exit 0 +fi + +default_branch() { + local ref branch + ref=$(git -C "$NAS_PATH" symbolic-ref --quiet --short refs/remotes/origin/HEAD 2>/dev/null || true) + if [ -n "$ref" ]; then + echo "${ref#origin/}" + return 0 + fi + for branch in main master; do + if git -C "$NAS_PATH" show-ref --verify --quiet "refs/heads/$branch"; then + echo "$branch" + return 0 + fi + done + return 1 +} + +DEFAULT=$(default_branch) || { + echo "$NAME: skipped: cannot determine default branch at $NAS_PATH" + exit 0 +} +BASE="origin/$DEFAULT" +if ! git -C "$NAS_PATH" rev-parse --verify --quiet "$BASE^{commit}" >/dev/null; then + echo "$NAME: skipped: $BASE does not exist at $NAS_PATH" + exit 0 +fi + +cur=$(git -C "$NAS_PATH" symbolic-ref --quiet --short HEAD 2>/dev/null || echo "") +dirty=no +[ -z "$(git -C "$NAS_PATH" status --porcelain 2>/dev/null | head -1)" ] || dirty=yes + +if [ "$cur" != "$DEFAULT" ]; then + echo "$NAME: STUCK: NAS checkout at $NAS_PATH is not on $DEFAULT - needs attention" + exit 0 +fi +if [ "$dirty" = yes ]; then + echo "$NAME: STUCK: NAS checkout at $NAS_PATH has uncommitted changes - needs attention" + exit 0 +fi + +local_rev="" +if ! local_rev=$(git -C "$NAS_PATH" rev-parse --verify --quiet "$DEFAULT"); then + echo "$NAME: skipped: cannot read local $DEFAULT at $NAS_PATH" + exit 0 +fi +remote_rev="" +if ! remote_rev=$(git -C "$NAS_PATH" rev-parse --verify --quiet "$BASE"); then + echo "$NAME: skipped: cannot read $BASE at $NAS_PATH" + exit 0 +fi +if [ "$local_rev" = "$remote_rev" ]; then + echo "$NAME: already current" + exit 0 +fi +if ! git -C "$NAS_PATH" merge-base --is-ancestor "$DEFAULT" "$BASE"; then + behind=$(git -C "$NAS_PATH" rev-list --count "HEAD..$BASE" 2>/dev/null) || behind="?" + echo "$NAME: STUCK: NAS checkout at $NAS_PATH diverged from $BASE, $behind commits behind - needs attention" + exit 0 +fi + +before=$(git -C "$NAS_PATH" rev-parse --short "$DEFAULT") +merge_output="" +if ! merge_output=$(git -C "$NAS_PATH" merge --ff-only "$BASE" 2>&1); then + reason="fast-forward failed" + [ -z "$merge_output" ] || reason="$reason: $(first_line "$merge_output")" + echo "$NAME: skipped: $reason" + exit 0 +fi +after=$(git -C "$NAS_PATH" rev-parse --short "$DEFAULT") + +# process_online : true when pm2 currently reports as online. +process_online() { + local proc=$1 status + status=$(pm2 jlist 2>/dev/null | jq -r --arg n "$proc" \ + '[.[] | select(.name == $n) | .pm2_env.status] | .[0] // "missing"' 2>/dev/null) || status="missing" + [ "$status" = "online" ] +} + +restart_problems="" +IFS=',' read -r -a procs <<< "$PROCS" +for proc in "${procs[@]}"; do + proc="${proc#"${proc%%[![:space:]]*}"}" + proc="${proc%"${proc##*[![:space:]]}"}" + [ -n "$proc" ] || continue + if ! pm2 restart "$proc" >/dev/null 2>&1; then + restart_problems="$restart_problems $proc(restart failed)" + continue + fi + if ! process_online "$proc"; then + restart_problems="$restart_problems $proc(not online after restart)" + fi +done + +if [ -n "$restart_problems" ]; then + echo "$NAME: synced $before..$after but restart verification failed:$restart_problems - needs attention" + exit 0 +fi +echo "$NAME: synced $before..$after, restarted $PROCS and verified online" diff --git a/bin/fm-teardown.sh b/bin/fm-teardown.sh index b59dd848d..d485b493e 100755 --- a/bin/fm-teardown.sh +++ b/bin/fm-teardown.sh @@ -1069,5 +1069,14 @@ rm -f "$STATE/$ID.status" "$STATE/$ID.turn-ended" "$STATE/$ID.check.sh" "$STATE/ if [ "$KIND" != scout ] && [ "$KIND" != secondmate ] && [ "$MODE" != local-only ]; then "$FM_ROOT/bin/fm-fleet-sync.sh" "$PROJ" || true fi +# Best-effort, non-blocking: a landed ship task (PR merged, or a local-only merge +# to local main) may have a live NAS deployment that also needs pulling and +# restarting - see bin/fm-nas-deploy-sync.sh's header. Scouts and secondmates +# never land a project change, so they are excluded the same as the fleet-sync +# call above; unlike that call, local-only IS included here, since a local-only +# merge is just as much a landed change as a PR merge. +if [ "$KIND" != scout ] && [ "$KIND" != secondmate ]; then + "$FM_ROOT/bin/fm-nas-deploy-sync.sh" "$(basename "$PROJ")" || true +fi echo "teardown $ID complete (window $T, worktree $WT)" backlog_refresh_reminder diff --git a/tests/fm-nas-deploy-sync.test.sh b/tests/fm-nas-deploy-sync.test.sh new file mode 100755 index 000000000..f16b9b170 --- /dev/null +++ b/tests/fm-nas-deploy-sync.test.sh @@ -0,0 +1,335 @@ +#!/usr/bin/env bash +# Behavior tests for fm-nas-deploy-sync.sh. +# +# A merged PR does not, by itself, reach the live site: every deployed app runs +# from its own separate checkout under /mnt/nas/experiments//, managed by +# pm2, independent from the projects/ clone firstmate/crewmates develop in +# (see the 2026-07-04 entry in data/learnings.md). fm-nas-deploy-sync.sh closes +# that gap: fast-forward the NAS checkout, restart and verify the pm2 +# process(es) that serve it - mirroring bin/fm-fleet-sync.sh's dirty/diverged +# safety exactly (never force, never touch a dirty or diverged checkout). +# +# This suite never touches a real /mnt/nas/experiments/ path or a real pm2 +# process: NAS checkouts are throwaway git repos under a per-test tmp home, and +# pm2 is a PATH-shimmed stub that records restarts and reports them online. +# +# Covers: +# - a clean, behind checkout fast-forwards and its pm2 process(es) restart and +# verify online +# - a dirty checkout is left completely untouched (reported, not fast-forwarded) +# - a diverged checkout is reported STUCK-style and left completely untouched +# - a project absent from data/nas-deployments.md (or the file itself absent) +# is a silent no-op: one skip line, exit 0 +# - bin/fm-teardown.sh's post-landing hook actually invokes the real script for +# a landed ship task, end to end (the NAS fixture fast-forwards and its pm2 +# process is restarted as a side effect of running teardown) +set -u + +# shellcheck source=tests/lib.sh disable=SC1091 +. "$(dirname "${BASH_SOURCE[0]}")/lib.sh" +fm_git_identity fmtest fmtest@example.invalid + +SYNC="$ROOT/bin/fm-nas-deploy-sync.sh" +TEARDOWN="$ROOT/bin/fm-teardown.sh" +TMP_ROOT=$(fm_test_tmproot fm-nas-deploy-sync-tests) +HOME_N=0 + +# --- generic fixtures -------------------------------------------------------- + +new_home() { + HOME_N=$((HOME_N + 1)) + local h="$TMP_ROOT/home-$HOME_N" + mkdir -p "$h/data" "$h/fakebin" + printf '%s\n' "$h" +} + +commit_file() { + local dir=$1 file=$2 content=$3 msg=$4 + printf '%s\n' "$content" > "$dir/$file" + git -C "$dir" add -- "$file" + git -C "$dir" commit -qm "$msg" +} + +# build_nas_pair : a bare "remote" origin plus a "work" repo wired to +# it (for advancing origin later) and a plain clone standing in for the NAS +# checkout, all with one commit on main. Echoes the NAS checkout path. +build_nas_pair() { + local home=$1 name=$2 work remote nas remote_abs + work="$home/work-$name" + remote="$home/remotes/$name.git" + nas="$home/nas-$name" + mkdir -p "$home/remotes" + + git init -q "$work" + git -C "$work" symbolic-ref HEAD refs/heads/main + commit_file "$work" file.txt v0 C0 + + git clone --quiet --bare "$work" "$remote" + remote_abs=$(cd "$remote" && pwd) + git -C "$work" remote add origin "file://$remote_abs" + git -C "$work" push -q -u origin main + + git clone --quiet "file://$remote_abs" "$nas" + printf '%s\n' "$nas" +} + +# advance_origin : push one more commit to 's origin via +# its work repo, so the NAS checkout (until synced) is behind origin/main. +advance_origin() { + local home=$1 name=$2 msg=$3 work + work="$home/work-$name" + commit_file "$work" file.txt "$msg" "$msg" + git -C "$work" push -q origin main +} + +head_sha() { git -C "$1" rev-parse HEAD; } + +# write_map : (re)write a one-row +# nas-deployments.md pipe table at $home/data/nas-deployments.md. +write_map() { + local home=$1 project=$2 nas_path=$3 procs=$4 + { + printf '| project | nas_repo_path | pm2_process(es) |\n' + printf '| --- | --- | --- |\n' + printf '| %s | %s | %s |\n' "$project" "$nas_path" "$procs" + } > "$home/data/nas-deployments.md" +} + +# write_pm2_stub : `pm2 restart ` appends to +# $restart_log and succeeds; `pm2 jlist` reports every logged name as online, so a +# process comes back "online" exactly when (and only when) it was restarted. +write_pm2_stub() { + local fakebin=$1 log=$2 + : > "$log" + cat > "$fakebin/pm2" <> "$log" + exit 0 + ;; + jlist) + printf '[' + first=1 + while IFS= read -r name; do + [ -n "\$name" ] || continue + [ "\$first" = 1 ] || printf ',' + first=0 + printf '{"name":"%s","pm2_env":{"status":"online"}}' "\$name" + done < "$log" + printf ']\n' + ;; + *) exit 0 ;; +esac +SH + chmod +x "$fakebin/pm2" +} + +# run_sync [args...]: run the sync script against an isolated home. +run_sync() { + local home=$1 + shift + FM_HOME="$home" PATH="$home/fakebin:$PATH" "$SYNC" "$@" +} + +# --- tests: standalone script ------------------------------------------------- + +test_clean_behind_fast_forwards_and_restarts() { + local home nas log out origin_head nas_head + home=$(new_home) + nas=$(build_nas_pair "$home" carscanner-test) + advance_origin "$home" carscanner-test C1 + write_map "$home" carscanner-test "$nas" carscanner-test + log="$home/restart.log" + write_pm2_stub "$home/fakebin" "$log" + + out=$(run_sync "$home" carscanner-test) + + assert_contains "$out" "carscanner-test: synced" "clean-behind: did not report a sync" + assert_contains "$out" "restarted carscanner-test and verified online" "clean-behind: did not report a verified restart" + origin_head=$(git -C "$home/work-carscanner-test" rev-parse main) + nas_head=$(head_sha "$nas") + [ "$origin_head" = "$nas_head" ] || fail "clean-behind: NAS checkout did not fast-forward to origin" + assert_grep carscanner-test "$log" "clean-behind: pm2 restart was not invoked" + pass "a clean, behind NAS checkout fast-forwards and its pm2 process restarts and verifies online" +} + +test_dirty_is_skipped_untouched() { + local home nas log out before after + home=$(new_home) + nas=$(build_nas_pair "$home" dirty-test) + advance_origin "$home" dirty-test C1 + before=$(head_sha "$nas") + printf 'uncommitted\n' >> "$nas/file.txt" + write_map "$home" dirty-test "$nas" dirty-test + log="$home/restart.log" + write_pm2_stub "$home/fakebin" "$log" + + out=$(run_sync "$home" dirty-test) + + assert_contains "$out" "STUCK" "dirty: did not report STUCK" + assert_contains "$out" "uncommitted changes" "dirty: did not name the dirty reason" + after=$(head_sha "$nas") + [ "$before" = "$after" ] || fail "dirty: NAS checkout HEAD moved despite being dirty" + git -C "$nas" diff --quiet -- file.txt && fail "dirty: uncommitted change was discarded" + [ ! -s "$log" ] || fail "dirty: pm2 restart was invoked on an untouched checkout" + pass "a dirty NAS checkout is reported STUCK and left completely untouched" +} + +test_diverged_is_stuck_untouched() { + local home nas log out before after + home=$(new_home) + nas=$(build_nas_pair "$home" diverged-test) + advance_origin "$home" diverged-test C1 + commit_file "$nas" local-only.txt local "local-only commit not on origin" + before=$(head_sha "$nas") + write_map "$home" diverged-test "$nas" diverged-test + log="$home/restart.log" + write_pm2_stub "$home/fakebin" "$log" + + out=$(run_sync "$home" diverged-test) + + assert_contains "$out" "STUCK" "diverged: did not report STUCK" + assert_contains "$out" "diverged" "diverged: did not name the diverged reason" + after=$(head_sha "$nas") + [ "$before" = "$after" ] || fail "diverged: NAS checkout HEAD moved despite diverging" + [ ! -s "$log" ] || fail "diverged: pm2 restart was invoked on an untouched checkout" + pass "a diverged NAS checkout is reported STUCK and left completely untouched" +} + +test_absent_project_is_silent_noop() { + local home nas out rc + home=$(new_home) + nas=$(build_nas_pair "$home" tracked-test) + write_map "$home" tracked-test "$nas" tracked-test + + set +e + out=$(run_sync "$home" untracked-project) + rc=$? + set -e + + expect_code 0 "$rc" "absent: sync should exit 0 for a project with no recorded deployment" + assert_contains "$out" "untracked-project: skipped: no recorded NAS deployment" "absent: did not report the expected skip line" + pass "a project absent from data/nas-deployments.md is a silent no-op" +} + +test_missing_map_file_is_silent_noop() { + local home out rc + home=$(new_home) + rm -f "$home/data/nas-deployments.md" + + set +e + out=$(run_sync "$home" anything) + rc=$? + set -e + + expect_code 0 "$rc" "no-map: sync should exit 0 when data/nas-deployments.md does not exist" + assert_contains "$out" "anything: skipped: no recorded NAS deployment" "no-map: did not report the expected skip line" + pass "a project checked with no data/nas-deployments.md at all is a silent no-op" +} + +# --- tests: teardown integration --------------------------------------------- + +# make_teardown_case : a trimmed fixture mirroring tests/fm-teardown.test.sh +# make_case's ALLOW path (no-mistakes mode, task branch pushed to origin), plus a +# NAS-fixture pair and a nas-deployments.md row for the project so the real +# post-teardown fm-nas-deploy-sync.sh call has real work to do. Echoes the case dir. +make_teardown_case() { + local name=$1 case_dir fakebin nas + case_dir="$TMP_ROOT/teardown-$name" + fakebin="$case_dir/fakebin" + mkdir -p "$case_dir/state" "$case_dir/config" "$case_dir/data" "$fakebin" + + cat > "$fakebin/treehouse" <<'SH' +#!/usr/bin/env bash +exit 0 +SH + cat > "$fakebin/tmux" <<'SH' +#!/usr/bin/env bash +exit 0 +SH + cat > "$fakebin/gh-axi" <<'SH' +#!/usr/bin/env bash +case "${1:-} ${2:-}" in + "pr list") printf '%s\n' "count: 0 (showing first 0)" "pull_requests[]: []" ; exit 0 ;; + "pr view") echo "error: pull request not found" >&2 ; exit 1 ;; +esac +exit 0 +SH + cat > "$fakebin/gh" <<'SH' +#!/usr/bin/env bash +case "${1:-} ${2:-}" in + "pr view") echo "error: pull request not found" >&2 ; exit 1 ;; +esac +exit 0 +SH + chmod +x "$fakebin/treehouse" "$fakebin/tmux" "$fakebin/gh-axi" "$fakebin/gh" + + git init -q --bare "$case_dir/origin.git" + git -C "$case_dir/origin.git" symbolic-ref HEAD refs/heads/main + git clone -q "$case_dir/origin.git" "$case_dir/_seed" 2>/dev/null + git -C "$case_dir/_seed" commit -q --allow-empty -m "origin baseline" + git -C "$case_dir/_seed" push -q origin main + rm -rf "$case_dir/_seed" + git clone -q "$case_dir/origin.git" "$case_dir/project" + git -C "$case_dir/project" remote set-head origin main 2>/dev/null || true + git -C "$case_dir/project" worktree add -q -b fm/task-x1 "$case_dir/wt" main + + touch "$case_dir/state/.last-watcher-beat" + + fm_write_meta "$case_dir/state/task-x1.meta" \ + "window=fm-task-x1" \ + "worktree=$case_dir/wt" \ + "project=$case_dir/project" \ + "kind=ship" \ + "mode=no-mistakes" + + git -C "$case_dir/wt" commit -q --allow-empty -m "shippable work" + git -C "$case_dir/wt" push -q origin fm/task-x1 + git -C "$case_dir/project" fetch -q origin + + nas=$(build_nas_pair "$case_dir" project-nas) + advance_origin "$case_dir" project-nas C1 + write_map "$case_dir" project "$nas" project-nas + write_pm2_stub "$fakebin" "$case_dir/restart.log" + + printf '%s\n' "$case_dir" +} + +run_teardown() { + local case_dir=$1 + FM_ROOT_OVERRIDE="$ROOT" \ + FM_STATE_OVERRIDE="$case_dir/state" \ + FM_CONFIG_OVERRIDE="$case_dir/config" \ + FM_NAS_DEPLOYMENTS_OVERRIDE="$case_dir/data/nas-deployments.md" \ + PATH="$case_dir/fakebin:$PATH" \ + "$TEARDOWN" task-x1 +} + +test_teardown_invokes_nas_deploy_sync_for_landed_project() { + local case_dir rc origin_head nas_head + case_dir=$(make_teardown_case invokes) + + set +e + run_teardown "$case_dir" > "$case_dir/stdout" 2> "$case_dir/stderr" + rc=$? + set -e + + expect_code 0 "$rc" "teardown-invokes: teardown should succeed for landed work" + ! grep -q REFUSED "$case_dir/stderr" || fail "teardown-invokes: teardown printed a REFUSED line" + + origin_head=$(git -C "$case_dir/work-project-nas" rev-parse main) + nas_head=$(head_sha "$case_dir/nas-project-nas") + [ "$origin_head" = "$nas_head" ] \ + || fail "teardown-invokes: post-teardown hook did not fast-forward the NAS fixture" + assert_grep project-nas "$case_dir/restart.log" \ + "teardown-invokes: post-teardown hook did not restart the NAS deployment's pm2 process" + pass "a landed ship-task teardown invokes fm-nas-deploy-sync.sh, which syncs and restarts the live NAS deployment" +} + +test_clean_behind_fast_forwards_and_restarts +test_dirty_is_skipped_untouched +test_diverged_is_stuck_untouched +test_absent_project_is_silent_noop +test_missing_map_file_is_silent_noop +test_teardown_invokes_nas_deploy_sync_for_landed_project From 38a5d73568cea8b7e659474c693c301318e55d95 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 20:31:13 -0400 Subject: [PATCH 27/41] no-mistakes(review): docs: document fm-nas-deploy-sync.sh script and override var --- docs/configuration.md | 1 + docs/scripts.md | 1 + 2 files changed, 2 insertions(+) diff --git a/docs/configuration.md b/docs/configuration.md index 1fcf64efe..f0da8fc90 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -403,6 +403,7 @@ FM_TIER_HEAVY_MIN_FILES=8 # fm-tier-guard.sh: changed-file count that e FM_WATCH_TRIAGE_LOG_MAX_BYTES=262144 # size cap for the watcher's absorbed-wake debug log FM_FLEET_SYNC_BOOTSTRAP_TIMEOUT= # optional seconds allowed for bootstrap's best-effort clone refresh; unset/blank defaults to max(20, 5 + 3 * origin-backed-project-count) FM_FLEET_PRUNE=1 # set to 0 to skip pruning local branches whose upstream is gone +FM_NAS_DEPLOYMENTS_OVERRIDE= # alternate data/nas-deployments.md path for fm-nas-deploy-sync.sh, mainly for tests FM_STALE_WORKTREE_LOCK_AGE_SECS=30 # min mtime age before fm-teardown.sh treats a leftover worktree git index.lock as provably stale FM_TREEHOUSE_RETURN_LOCK_RETRIES=3 # retries after a treehouse return fails on the transient git index.lock signature FM_TREEHOUSE_RETURN_LOCK_RETRY_WAIT_SECS=1 # seconds fm-teardown.sh waits before each retry after that signature diff --git a/docs/scripts.md b/docs/scripts.md index 71bfa3e1b..99e342184 100644 --- a/docs/scripts.md +++ b/docs/scripts.md @@ -69,6 +69,7 @@ The shared no-mistakes gate refusal used by `fm-spawn.sh`, `fm-send.sh`, and `fm | `fm-pr-merge.sh` | Record PR metadata, then merge a task's PR from its full GitHub URL | | `fm-promote.sh` | Promote a scout task in place to a protected ship task | | `fm-teardown.sh` | Fail-closed teardown: return landed ship worktrees, require scout reports, retire secondmate homes | +| `fm-nas-deploy-sync.sh` | Best-effort, non-fatal fast-forward sync and pm2 restart of a project's live NAS checkout after a landed ship-task teardown, mirroring `fm-fleet-sync.sh`'s dirty/diverged safety | | `fm-harness.sh` | Detect the running harness and resolve crew or secondmate harness, model, and effort | | `fm-lock.sh` | Per-home firstmate session lock | | `fm-x-lib.sh` | Shared X-mode config, relay, and reply-threading helpers | From ccac7cf9cb1cea657271709135da4da99a4a37f8 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 21:46:31 -0400 Subject: [PATCH 28/41] no-mistakes(review): Sandbox teardown tests' NAS deploy sync lookup file --- tests/fm-teardown.test.sh | 1 + 1 file changed, 1 insertion(+) diff --git a/tests/fm-teardown.test.sh b/tests/fm-teardown.test.sh index bf9dbd7fc..b49a5d778 100755 --- a/tests/fm-teardown.test.sh +++ b/tests/fm-teardown.test.sh @@ -526,6 +526,7 @@ run_teardown() { FM_ROOT_OVERRIDE="$ROOT" \ FM_STATE_OVERRIDE="$case_dir/state" \ FM_CONFIG_OVERRIDE="$case_dir/config" \ + FM_NAS_DEPLOYMENTS_OVERRIDE="$case_dir/data/nas-deployments.md" \ PATH="$case_dir/fakebin:$PATH" \ "$TEARDOWN" task-x1 "$@" } From d5f69594b2df33dd08d34c6ba75ca5530931f814 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 22:00:06 -0400 Subject: [PATCH 29/41] no-mistakes(review): Aggregate pm2 cluster status and stub NAS sync in gotmp tests --- bin/fm-nas-deploy-sync.sh | 10 ++++++-- tests/fm-gotmp.test.sh | 12 +++++++++ tests/fm-nas-deploy-sync.test.sh | 44 ++++++++++++++++++++++++++++++++ 3 files changed, 64 insertions(+), 2 deletions(-) diff --git a/bin/fm-nas-deploy-sync.sh b/bin/fm-nas-deploy-sync.sh index 8899703d0..0608abfb5 100755 --- a/bin/fm-nas-deploy-sync.sh +++ b/bin/fm-nas-deploy-sync.sh @@ -161,11 +161,17 @@ if ! merge_output=$(git -C "$NAS_PATH" merge --ff-only "$BASE" 2>&1); then fi after=$(git -C "$NAS_PATH" rev-parse --short "$DEFAULT") -# process_online : true when pm2 currently reports as online. +# process_online : true when pm2 currently reports EVERY instance of +# as online (a clustered app can register several jlist entries under +# the same name; one lagging instance must not be masked by the others). process_online() { local proc=$1 status status=$(pm2 jlist 2>/dev/null | jq -r --arg n "$proc" \ - '[.[] | select(.name == $n) | .pm2_env.status] | .[0] // "missing"' 2>/dev/null) || status="missing" + '[.[] | select(.name == $n) | .pm2_env.status] as $s + | if ($s | length) == 0 then "missing" + elif all($s[]; . == "online") then "online" + else "not-online" + end' 2>/dev/null) || status="missing" [ "$status" = "online" ] } diff --git a/tests/fm-gotmp.test.sh b/tests/fm-gotmp.test.sh index 6b6fd8aa1..2a4821da9 100755 --- a/tests/fm-gotmp.test.sh +++ b/tests/fm-gotmp.test.sh @@ -75,6 +75,13 @@ SH exit 0 SH chmod +x "$fake/bin/fm-fleet-sync.sh" + # fm-nas-deploy-sync.sh: stub (called for the same non-scout/non-local-only + # teardowns, right after fm-fleet-sync.sh). + cat > "$fake/bin/fm-nas-deploy-sync.sh" <<'SH' +#!/usr/bin/env bash +exit 0 +SH + chmod +x "$fake/bin/fm-nas-deploy-sync.sh" # fm-tasks-axi-lib.sh: stub (teardown sources it). Report no backend so # backlog_refresh_reminder takes the plain-message path; no tasks-axi here. cat > "$fake/bin/fm-tasks-axi-lib.sh" <<'SH' @@ -171,6 +178,11 @@ SH exit 0 SH chmod +x "$fake/bin/fm-fleet-sync.sh" + cat > "$fake/bin/fm-nas-deploy-sync.sh" <<'SH' +#!/usr/bin/env bash +exit 0 +SH + chmod +x "$fake/bin/fm-nas-deploy-sync.sh" cat > "$fake/bin/fm-tasks-axi-lib.sh" <<'SH' fm_tasks_axi_backend_available() { return 1; } SH diff --git a/tests/fm-nas-deploy-sync.test.sh b/tests/fm-nas-deploy-sync.test.sh index f16b9b170..4870148c3 100755 --- a/tests/fm-nas-deploy-sync.test.sh +++ b/tests/fm-nas-deploy-sync.test.sh @@ -125,6 +125,29 @@ SH chmod +x "$fakebin/pm2" } +# write_pm2_cluster_stub : like write_pm2_stub, but +# `pm2 jlist` always reports TWO instances of - one online, one stopped - +# regardless of what was restarted, simulating a clustered pm2 app where one +# instance in the cluster never comes back after a restart. +write_pm2_cluster_stub() { + local fakebin=$1 log=$2 proc=$3 + : > "$log" + cat > "$fakebin/pm2" <> "$log" + exit 0 + ;; + jlist) + printf '[{"name":"%s","pm2_env":{"status":"online"}},{"name":"%s","pm2_env":{"status":"stopped"}}]\n' "$proc" "$proc" + ;; + *) exit 0 ;; +esac +SH + chmod +x "$fakebin/pm2" +} + # run_sync [args...]: run the sync script against an isolated home. run_sync() { local home=$1 @@ -213,6 +236,26 @@ test_absent_project_is_silent_noop() { pass "a project absent from data/nas-deployments.md is a silent no-op" } +test_clustered_partial_restart_is_not_masked_online() { + local home nas log out + home=$(new_home) + nas=$(build_nas_pair "$home" cluster-test) + advance_origin "$home" cluster-test C1 + write_map "$home" cluster-test "$nas" cluster-test + log="$home/restart.log" + write_pm2_cluster_stub "$home/fakebin" "$log" cluster-test + + out=$(run_sync "$home" cluster-test) + + assert_contains "$out" "not online after restart" \ + "clustered: did not report the lagging cluster instance as a restart problem" + assert_contains "$out" "needs attention" "clustered: did not flag the deployment as needing attention" + case "$out" in + *"verified online"*) fail "clustered: a lagging cluster instance was masked as fully verified online" ;; + esac + pass "a clustered pm2 app with one lagging instance is reported as a restart problem, not masked as online" +} + test_missing_map_file_is_silent_noop() { local home out rc home=$(new_home) @@ -330,6 +373,7 @@ test_teardown_invokes_nas_deploy_sync_for_landed_project() { test_clean_behind_fast_forwards_and_restarts test_dirty_is_skipped_untouched test_diverged_is_stuck_untouched +test_clustered_partial_restart_is_not_masked_online test_absent_project_is_silent_noop test_missing_map_file_is_silent_noop test_teardown_invokes_nas_deploy_sync_for_landed_project From 5e577d1b000bbf483a84e92d382f4b20ffb90b16 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 22:15:31 -0400 Subject: [PATCH 30/41] no-mistakes(review): Bound NAS deploy sync git/fs calls with a timeout --- bin/fm-nas-deploy-sync.sh | 61 +++++++++++++++++++++++++++++---------- 1 file changed, 45 insertions(+), 16 deletions(-) diff --git a/bin/fm-nas-deploy-sync.sh b/bin/fm-nas-deploy-sync.sh index 0608abfb5..d8d47c4d6 100755 --- a/bin/fm-nas-deploy-sync.sh +++ b/bin/fm-nas-deploy-sync.sh @@ -27,6 +27,12 @@ # an alternate mapping file directly. pm2 is invoked via PATH, so a fakebin shim # ahead of it on PATH intercepts restart/list during tests - real tests must never # reach the real /mnt/nas/experiments or a real pm2 process. +# +# Every filesystem/git touch of $NAS_PATH is wrapped in a bounded timeout +# (FM_NAS_SYNC_TIMEOUT seconds, default 15): the mount is a NAS, and an +# unreachable one is a hang risk (a stuck stat/read), not a fast failure, which +# would otherwise stall fm-teardown.sh's caller and break this script's +# non-blocking contract. set -eu SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" @@ -34,6 +40,29 @@ FM_ROOT="${FM_ROOT_OVERRIDE:-$(cd "$SCRIPT_DIR/.." && pwd)}" FM_HOME="${FM_HOME:-${FM_ROOT_OVERRIDE:-$FM_ROOT}}" DATA="${FM_DATA_OVERRIDE:-$FM_HOME/data}" MAP="${FM_NAS_DEPLOYMENTS_OVERRIDE:-$DATA/nas-deployments.md}" +NAS_SYNC_TIMEOUT="${FM_NAS_SYNC_TIMEOUT:-15}" + +HAVE_TIMEOUT=none +if command -v timeout >/dev/null 2>&1; then + HAVE_TIMEOUT=timeout +elif command -v gtimeout >/dev/null 2>&1; then + HAVE_TIMEOUT=gtimeout +fi + +# bounded : run , killed after $NAS_SYNC_TIMEOUT seconds when +# a `timeout`/`gtimeout` binary is available, run directly otherwise. +bounded() { + case "$HAVE_TIMEOUT" in + timeout) timeout "$NAS_SYNC_TIMEOUT" "$@" ;; + gtimeout) gtimeout "$NAS_SYNC_TIMEOUT" "$@" ;; + *) "$@" ;; + esac +} + +# git_nas : git -C "$NAS_PATH" , bounded. +git_nas() { + bounded git -C "$NAS_PATH" "$@" +} usage() { echo "usage: fm-nas-deploy-sync.sh " >&2 @@ -71,21 +100,21 @@ fi NAS_PATH=${row%%|*} PROCS=${row#*|} -if [ ! -d "$NAS_PATH" ]; then +if ! bounded test -d "$NAS_PATH"; then echo "$NAME: skipped: NAS path $NAS_PATH not a directory" exit 0 fi -if ! git -C "$NAS_PATH" rev-parse --is-inside-work-tree >/dev/null 2>&1; then +if ! git_nas rev-parse --is-inside-work-tree >/dev/null 2>&1; then echo "$NAME: skipped: $NAS_PATH not a git repo" exit 0 fi -if ! git -C "$NAS_PATH" remote get-url origin >/dev/null 2>&1; then +if ! git_nas remote get-url origin >/dev/null 2>&1; then echo "$NAME: skipped: no origin remote at $NAS_PATH" exit 0 fi fetch_output="" -if ! fetch_output=$(git -C "$NAS_PATH" fetch origin --quiet 2>&1); then +if ! fetch_output=$(git_nas fetch origin --quiet 2>&1); then reason="fetch failed" [ -z "$fetch_output" ] || reason="$reason: $(first_line "$fetch_output")" echo "$NAME: skipped: $reason" @@ -94,13 +123,13 @@ fi default_branch() { local ref branch - ref=$(git -C "$NAS_PATH" symbolic-ref --quiet --short refs/remotes/origin/HEAD 2>/dev/null || true) + ref=$(git_nas symbolic-ref --quiet --short refs/remotes/origin/HEAD 2>/dev/null || true) if [ -n "$ref" ]; then echo "${ref#origin/}" return 0 fi for branch in main master; do - if git -C "$NAS_PATH" show-ref --verify --quiet "refs/heads/$branch"; then + if git_nas show-ref --verify --quiet "refs/heads/$branch"; then echo "$branch" return 0 fi @@ -113,14 +142,14 @@ DEFAULT=$(default_branch) || { exit 0 } BASE="origin/$DEFAULT" -if ! git -C "$NAS_PATH" rev-parse --verify --quiet "$BASE^{commit}" >/dev/null; then +if ! git_nas rev-parse --verify --quiet "$BASE^{commit}" >/dev/null; then echo "$NAME: skipped: $BASE does not exist at $NAS_PATH" exit 0 fi -cur=$(git -C "$NAS_PATH" symbolic-ref --quiet --short HEAD 2>/dev/null || echo "") +cur=$(git_nas symbolic-ref --quiet --short HEAD 2>/dev/null || echo "") dirty=no -[ -z "$(git -C "$NAS_PATH" status --porcelain 2>/dev/null | head -1)" ] || dirty=yes +[ -z "$(git_nas status --porcelain 2>/dev/null | head -1)" ] || dirty=yes if [ "$cur" != "$DEFAULT" ]; then echo "$NAME: STUCK: NAS checkout at $NAS_PATH is not on $DEFAULT - needs attention" @@ -132,12 +161,12 @@ if [ "$dirty" = yes ]; then fi local_rev="" -if ! local_rev=$(git -C "$NAS_PATH" rev-parse --verify --quiet "$DEFAULT"); then +if ! local_rev=$(git_nas rev-parse --verify --quiet "$DEFAULT"); then echo "$NAME: skipped: cannot read local $DEFAULT at $NAS_PATH" exit 0 fi remote_rev="" -if ! remote_rev=$(git -C "$NAS_PATH" rev-parse --verify --quiet "$BASE"); then +if ! remote_rev=$(git_nas rev-parse --verify --quiet "$BASE"); then echo "$NAME: skipped: cannot read $BASE at $NAS_PATH" exit 0 fi @@ -145,21 +174,21 @@ if [ "$local_rev" = "$remote_rev" ]; then echo "$NAME: already current" exit 0 fi -if ! git -C "$NAS_PATH" merge-base --is-ancestor "$DEFAULT" "$BASE"; then - behind=$(git -C "$NAS_PATH" rev-list --count "HEAD..$BASE" 2>/dev/null) || behind="?" +if ! git_nas merge-base --is-ancestor "$DEFAULT" "$BASE"; then + behind=$(git_nas rev-list --count "HEAD..$BASE" 2>/dev/null) || behind="?" echo "$NAME: STUCK: NAS checkout at $NAS_PATH diverged from $BASE, $behind commits behind - needs attention" exit 0 fi -before=$(git -C "$NAS_PATH" rev-parse --short "$DEFAULT") +before=$(git_nas rev-parse --short "$DEFAULT") merge_output="" -if ! merge_output=$(git -C "$NAS_PATH" merge --ff-only "$BASE" 2>&1); then +if ! merge_output=$(git_nas merge --ff-only "$BASE" 2>&1); then reason="fast-forward failed" [ -z "$merge_output" ] || reason="$reason: $(first_line "$merge_output")" echo "$NAME: skipped: $reason" exit 0 fi -after=$(git -C "$NAS_PATH" rev-parse --short "$DEFAULT") +after=$(git_nas rev-parse --short "$DEFAULT") # process_online : true when pm2 currently reports EVERY instance of # as online (a clustered app can register several jlist entries under From dcdf3505b2cb4b0ede05f40355d2c2077fbfd83b Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 22:41:42 -0400 Subject: [PATCH 31/41] no-mistakes(review): Guard NAS sync fetch race, document timeout knob, add tests --- bin/fm-nas-deploy-sync.sh | 96 +++++++++++++++++++++++++++++++- docs/configuration.md | 4 ++ tests/fm-nas-deploy-sync.test.sh | 94 +++++++++++++++++++++++++++++++ 3 files changed, 191 insertions(+), 3 deletions(-) diff --git a/bin/fm-nas-deploy-sync.sh b/bin/fm-nas-deploy-sync.sh index d8d47c4d6..1233dc9f7 100755 --- a/bin/fm-nas-deploy-sync.sh +++ b/bin/fm-nas-deploy-sync.sh @@ -33,6 +33,15 @@ # unreachable one is a hang risk (a stuck stat/read), not a fast failure, which # would otherwise stall fm-teardown.sh's caller and break this script's # non-blocking contract. +# +# $NAS_PATH is one single shared live-deployment checkout - unlike a project's +# projects/ dev clone, which is scoped to one firstmate home, two ship +# tasks for the same project landing moments apart can both reach the SAME +# NAS checkout from concurrent teardowns. That races the fetch on an orphaned +# .git/packed-refs.lock exactly like bin/fm-fleet-sync.sh's fetch had to guard +# against; fetch_with_packed_refs_lock_guard below mirrors that recovery, +# sharing the staleness proof from bin/fm-lock-lib.sh, bounded by +# FM_NAS_SYNC_PACKED_REFS_LOCK_RETRIES / _RETRY_WAIT_SECS / _AGE_SECS. set -eu SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)" @@ -41,6 +50,19 @@ FM_HOME="${FM_HOME:-${FM_ROOT_OVERRIDE:-$FM_ROOT}}" DATA="${FM_DATA_OVERRIDE:-$FM_HOME/data}" MAP="${FM_NAS_DEPLOYMENTS_OVERRIDE:-$DATA/nas-deployments.md}" NAS_SYNC_TIMEOUT="${FM_NAS_SYNC_TIMEOUT:-15}" +# shellcheck source=bin/fm-lock-lib.sh +. "$SCRIPT_DIR/fm-lock-lib.sh" +FM_LOCK_LOG_PREFIX=nas-deploy-sync + +NAS_SYNC_PACKED_REFS_LOCK_RETRIES=${FM_NAS_SYNC_PACKED_REFS_LOCK_RETRIES:-3} +NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS=${FM_NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS:-1} +NAS_SYNC_PACKED_REFS_LOCK_AGE_SECS=${FM_NAS_SYNC_PACKED_REFS_LOCK_AGE_SECS:-30} +case "$NAS_SYNC_PACKED_REFS_LOCK_RETRIES" in ''|*[!0-9]*) NAS_SYNC_PACKED_REFS_LOCK_RETRIES=3 ;; esac +case "$NAS_SYNC_PACKED_REFS_LOCK_AGE_SECS" in ''|*[!0-9]*) NAS_SYNC_PACKED_REFS_LOCK_AGE_SECS=30 ;; esac +if ! [[ "$NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS" =~ ^([0-9]+([.][0-9]*)?|[.][0-9]+)$ ]]; then + echo "nas-deploy-sync: invalid packed-refs lock retry wait '$NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS'; using 1s" >&2 + NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS=1 +fi HAVE_TIMEOUT=none if command -v timeout >/dev/null 2>&1; then @@ -79,6 +101,75 @@ first_line() { printf '%s\n' "$1" | sed -n '1s/[[:space:]]\{1,\}/ /g;1p' } +# True when git stderr shows the packed-refs.lock "File exists" race, mirroring +# bin/fm-fleet-sync.sh's is_packed_refs_lock_error. +is_packed_refs_lock_error() { + printf '%s\n' "$1" | grep -Eq "Unable to create ['\"].*packed-refs\\.lock['\"]: File exists" +} + +# Absolute path to $NAS_PATH's packed-refs.lock, or empty when it cannot be +# resolved. +packed_refs_lock_path() { + local lock abs + lock=$(git_nas rev-parse --git-path packed-refs.lock 2>/dev/null) || return 1 + [ -n "$lock" ] || return 1 + case "$lock" in + /*) printf '%s\n' "$lock" ;; + *) + abs=$(cd "$NAS_PATH" && pwd -P) || return 1 + printf '%s/%s\n' "$abs" "$lock" + ;; + esac +} + +# Run `git_nas fetch origin --quiet`, tolerating an orphaned packed-refs.lock +# left by a second landed-teardown racing this exact NAS checkout. Sets +# FETCH_OUTPUT and returns the fetch's exit status; mirrors bin/fm-fleet-sync.sh's +# fetch_with_packed_refs_lock_guard retry-then-provably-stale-clear contract, +# sharing the staleness proof from bin/fm-lock-lib.sh. +fetch_with_packed_refs_lock_guard() { + local rc attempt=0 lock lock_desc + FETCH_OUTPUT=$(git_nas fetch origin --quiet 2>&1); rc=$? + [ "$rc" -eq 0 ] && return 0 + is_packed_refs_lock_error "$FETCH_OUTPUT" || return "$rc" + + lock=$(packed_refs_lock_path) || lock="" + lock_desc=${lock:-packed-refs.lock} + while [ "$attempt" -lt "$NAS_SYNC_PACKED_REFS_LOCK_RETRIES" ]; do + attempt=$(( attempt + 1 )) + echo "$NAME: fetch blocked by packed-refs lock ($lock_desc) at $NAS_PATH; waiting ${NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS}s and retrying ($attempt/${NAS_SYNC_PACKED_REFS_LOCK_RETRIES}) (owning process may be exiting)" >&2 + sleep "$NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS" + FETCH_OUTPUT=$(git_nas fetch origin --quiet 2>&1); rc=$? + if [ "$rc" -eq 0 ]; then + echo "$NAME: fetch succeeded on retry; packed-refs lock cleared on its own" >&2 + return 0 + fi + is_packed_refs_lock_error "$FETCH_OUTPUT" || return "$rc" + done + + # Retries exhausted and still the lock signature. Clear ONLY if provably stale. + lock=$(packed_refs_lock_path) || lock="" + if [ -n "$lock" ] && [ -e "$lock" ]; then + if fm_lock_is_provably_stale "$lock" "$NAS_PATH" "$NAS_SYNC_PACKED_REFS_LOCK_AGE_SECS"; then + if ! rm -f "$lock"; then + echo "$NAME: failed to remove provably-stale packed-refs lock $lock; leaving it in place" >&2 + return "$rc" + fi + echo "$NAME: removed provably-stale packed-refs lock $lock (age >= ${NAS_SYNC_PACKED_REFS_LOCK_AGE_SECS}s, no live holder) and retrying fetch" >&2 + FETCH_OUTPUT=$(git_nas fetch origin --quiet 2>&1); rc=$? + if [ "$rc" -eq 0 ]; then + echo "$NAME: fetch succeeded after stale packed-refs lock cleanup" >&2 + return 0 + fi + return "$rc" + fi + echo "$NAME: fetch blocked by packed-refs lock $lock that persisted across ${NAS_SYNC_PACKED_REFS_LOCK_RETRIES} retries and is not provably stale (may belong to a live process); leaving it in place" >&2 + return "$rc" + fi + echo "$NAME: fetch packed-refs lock signature persisted across ${NAS_SYNC_PACKED_REFS_LOCK_RETRIES} retries even after the lock file disappeared" >&2 + return "$rc" +} + # lookup_row: print "|" for $NAME's row in $MAP, or # nothing when the file is absent or has no matching row. lookup_row() { @@ -113,10 +204,9 @@ if ! git_nas remote get-url origin >/dev/null 2>&1; then exit 0 fi -fetch_output="" -if ! fetch_output=$(git_nas fetch origin --quiet 2>&1); then +if ! fetch_with_packed_refs_lock_guard; then reason="fetch failed" - [ -z "$fetch_output" ] || reason="$reason: $(first_line "$fetch_output")" + [ -z "$FETCH_OUTPUT" ] || reason="$reason: $(first_line "$FETCH_OUTPUT")" echo "$NAME: skipped: $reason" exit 0 fi diff --git a/docs/configuration.md b/docs/configuration.md index f0da8fc90..bd787569a 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -404,6 +404,10 @@ FM_WATCH_TRIAGE_LOG_MAX_BYTES=262144 # size cap for the watcher's absorbed-wak FM_FLEET_SYNC_BOOTSTRAP_TIMEOUT= # optional seconds allowed for bootstrap's best-effort clone refresh; unset/blank defaults to max(20, 5 + 3 * origin-backed-project-count) FM_FLEET_PRUNE=1 # set to 0 to skip pruning local branches whose upstream is gone FM_NAS_DEPLOYMENTS_OVERRIDE= # alternate data/nas-deployments.md path for fm-nas-deploy-sync.sh, mainly for tests +FM_NAS_SYNC_TIMEOUT=15 # seconds allowed per filesystem/git touch of fm-nas-deploy-sync.sh's $NAS_PATH before it is killed, so an unreachable NAS mount cannot hang fm-teardown.sh's caller +FM_NAS_SYNC_PACKED_REFS_LOCK_RETRIES=3 # fetch retries after fm-nas-deploy-sync.sh hits the orphaned .git/packed-refs.lock signature on the shared NAS checkout (two landed teardowns for the same project can race it) +FM_NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS=1 # seconds fm-nas-deploy-sync.sh waits before each of those retries +FM_NAS_SYNC_PACKED_REFS_LOCK_AGE_SECS=30 # min mtime age before fm-nas-deploy-sync.sh treats a leftover packed-refs.lock as provably stale FM_STALE_WORKTREE_LOCK_AGE_SECS=30 # min mtime age before fm-teardown.sh treats a leftover worktree git index.lock as provably stale FM_TREEHOUSE_RETURN_LOCK_RETRIES=3 # retries after a treehouse return fails on the transient git index.lock signature FM_TREEHOUSE_RETURN_LOCK_RETRY_WAIT_SECS=1 # seconds fm-teardown.sh waits before each retry after that signature diff --git a/tests/fm-nas-deploy-sync.test.sh b/tests/fm-nas-deploy-sync.test.sh index 4870148c3..ecffe6e29 100755 --- a/tests/fm-nas-deploy-sync.test.sh +++ b/tests/fm-nas-deploy-sync.test.sh @@ -155,6 +155,53 @@ run_sync() { FM_HOME="$home" PATH="$home/fakebin:$PATH" "$SYNC" "$@" } +# write_git_hung_fetch_stub : any `fetch` call hangs indefinitely +# (well past FM_NAS_SYNC_TIMEOUT), so only the bounded wrapper's `timeout` kill +# can make the sync return promptly; every other call delegates to the real git. +write_git_hung_fetch_stub() { + local fakebin=$1 realgit + realgit=$(command -v git) + cat > "$fakebin/git" < : fail the FIRST +# `fetch` with the packed-refs.lock "File exists" signature, then delegate every +# later call - including the retried fetch - to the real git. Mirrors +# tests/fm-fleet-sync.test.sh's git_transient_packed_refs_lock, adapted to fake +# the stderr signature directly since fm-nas-deploy-sync.sh's fetch has no +# --prune to organically trigger a real packed-refs rewrite. +write_git_transient_packed_refs_lock_stub() { + local fakebin=$1 counter=$2 realgit + realgit=$(command -v git) + cat > "$fakebin/git" </dev/null || echo 0); n=\$(( n + 1 )) + printf '%s\n' "\$n" > "$counter" + if [ "\$n" -eq 1 ]; then + echo "error: could not delete reference refs/remotes/origin/feature: Unable to create '.git/packed-refs.lock': File exists." >&2 + exit 1 + fi +fi +exec "\$real" "\$@" +SH + chmod +x "$fakebin/git" +} + # --- tests: standalone script ------------------------------------------------- test_clean_behind_fast_forwards_and_restarts() { @@ -256,6 +303,51 @@ test_clustered_partial_restart_is_not_masked_online() { pass "a clustered pm2 app with one lagging instance is reported as a restart problem, not masked as online" } +test_hung_fetch_is_bounded_by_timeout() { + local home nas out rc start elapsed + if ! command -v timeout >/dev/null 2>&1 && ! command -v gtimeout >/dev/null 2>&1; then + pass "SKIP (no timeout/gtimeout binary): hung-fetch timeout bound check" + return + fi + home=$(new_home) + nas=$(build_nas_pair "$home" hungfetch-test) + advance_origin "$home" hungfetch-test C1 + write_map "$home" hungfetch-test "$nas" hungfetch-test + write_pm2_stub "$home/fakebin" "$home/restart.log" + write_git_hung_fetch_stub "$home/fakebin" + + start=$SECONDS + set +e + out=$(FM_NAS_SYNC_TIMEOUT=1 run_sync "$home" hungfetch-test 2>/dev/null) + rc=$? + set -e + elapsed=$(( SECONDS - start )) + + expect_code 0 "$rc" "hung-fetch: sync should still exit 0 when the fetch hangs" + [ "$elapsed" -lt 10 ] || fail "hung-fetch: sync took ${elapsed}s - the timeout bound did not stop the hung fetch (stub sleeps 20s)" + assert_contains "$out" "hungfetch-test: skipped: fetch failed" "hung-fetch: did not report the bounded fetch as a failure" + pass "a hung NAS fetch is killed by FM_NAS_SYNC_TIMEOUT instead of blocking the caller" +} + +test_transient_packed_refs_lock_is_retried() { + local home nas out counter err + home=$(new_home) + nas=$(build_nas_pair "$home" locktrans-test) + advance_origin "$home" locktrans-test C1 + write_map "$home" locktrans-test "$nas" locktrans-test + write_pm2_stub "$home/fakebin" "$home/restart.log" + counter="$home/git-fetch-count"; : > "$counter" + write_git_transient_packed_refs_lock_stub "$home/fakebin" "$counter" + err="$home/err-locktrans" + + out=$(FM_NAS_SYNC_PACKED_REFS_LOCK_RETRIES=3 FM_NAS_SYNC_PACKED_REFS_LOCK_RETRY_WAIT_SECS=0 \ + run_sync "$home" locktrans-test 2>"$err") + + assert_contains "$out" "locktrans-test: synced" "transient lock: sync did not complete after the lock self-cleared" + assert_grep "cleared on its own" "$err" "transient lock: guard did not report the self-clear" + pass "a transient packed-refs.lock signature on the NAS fetch is retried instead of giving up immediately" +} + test_missing_map_file_is_silent_noop() { local home out rc home=$(new_home) @@ -375,5 +467,7 @@ test_dirty_is_skipped_untouched test_diverged_is_stuck_untouched test_clustered_partial_restart_is_not_masked_online test_absent_project_is_silent_noop +test_hung_fetch_is_bounded_by_timeout +test_transient_packed_refs_lock_is_retried test_missing_map_file_is_silent_noop test_teardown_invokes_nas_deploy_sync_for_landed_project From 19897ca5456ec301b2bad0a0b916fa5149f82c08 Mon Sep 17 00:00:00 2001 From: Michael Date: Sat, 11 Jul 2026 23:49:03 -0400 Subject: [PATCH 32/41] no-mistakes(test): placeholder - awaiting background test run completion --- tests/fm-backlog-handoff.test.sh | 7 +++++-- tests/fm-secondmate-lifecycle-e2e.test.sh | 11 +++++++---- tests/fm-teardown.test.sh | 20 ++++++++++++++++++++ 3 files changed, 32 insertions(+), 6 deletions(-) diff --git a/tests/fm-backlog-handoff.test.sh b/tests/fm-backlog-handoff.test.sh index 9e10703e8..263ddad9c 100755 --- a/tests/fm-backlog-handoff.test.sh +++ b/tests/fm-backlog-handoff.test.sh @@ -8,10 +8,13 @@ set -u # shellcheck source=tests/secondmate-helpers.sh disable=SC1091 . "$(dirname "${BASH_SOURCE[0]}")/secondmate-helpers.sh" +# shellcheck source=bin/fm-tasks-axi-lib.sh disable=SC1091 +. "$ROOT/bin/fm-tasks-axi-lib.sh" # The move is delegated to `tasks-axi mv`, so this suite exercises the real -# binary. Skip cleanly when it is absent (matching the backend smoke suites). -command -v tasks-axi >/dev/null 2>&1 || { echo "skip: tasks-axi not found (required by the delegated handoff path)"; exit 0; } +# binary. Skip cleanly when it is absent or too old for the atomic multi-ID +# move the delegated handoff path requires (matching the backend smoke suites). +fm_tasks_axi_compatible || { echo "skip: tasks-axi not found or too old (required by the delegated handoff path)"; exit 0; } TMP_ROOT=$(fm_test_tmproot fm-backlog-handoff) diff --git a/tests/fm-secondmate-lifecycle-e2e.test.sh b/tests/fm-secondmate-lifecycle-e2e.test.sh index 099dd2682..3046df7e0 100755 --- a/tests/fm-secondmate-lifecycle-e2e.test.sh +++ b/tests/fm-secondmate-lifecycle-e2e.test.sh @@ -155,10 +155,13 @@ phase_send() { } phase_handoff() { - # The move is delegated to `tasks-axi mv`; skip cleanly when it is absent (the - # downstream recovery and teardown phases do not depend on this phase). - if ! command -v tasks-axi >/dev/null 2>&1; then - echo "skip: tasks-axi not found (backlog handoff delegates to it)" + # The move is delegated to `tasks-axi mv`; skip cleanly when it is absent or + # too old for atomic multi-ID moves (the downstream recovery and teardown + # phases do not depend on this phase). + # shellcheck source=bin/fm-tasks-axi-lib.sh disable=SC1091 + . "$ROOT/bin/fm-tasks-axi-lib.sh" + if ! fm_tasks_axi_compatible; then + echo "skip: tasks-axi not found or too old (backlog handoff delegates to it)" return 0 fi cat > "$HOME_DIR/data/backlog.md" <<'EOF' diff --git a/tests/fm-teardown.test.sh b/tests/fm-teardown.test.sh index b49a5d778..bfe0f8ee4 100755 --- a/tests/fm-teardown.test.sh +++ b/tests/fm-teardown.test.sh @@ -209,6 +209,18 @@ land_on_origin_main() { rm -rf "$tmp" } +# content_in_default (bin/fm-teardown.sh) relies on `git merge-tree +# --write-tree`, added in git 2.38. Probe the REAL capability rather than +# guessing a version cutoff: older git treats the flag as an unrecognized +# revision ("unknown rev --write-tree"), which the two tests exercising that +# fallback below cannot pass under. +fm_git_supports_merge_tree_write_tree() { + case "$(git -C "$ROOT" merge-tree --write-tree HEAD HEAD 2>&1)" in + *"unknown rev"*) return 1 ;; + *) return 0 ;; + esac +} + # Override GitHub lookups to report PR 7 as merged with the supplied head. add_gh_pr_merged_for_head() { local case_dir=$1 head=$2 @@ -899,6 +911,10 @@ test_pr_check_records_remote_head_when_local_lags() { test_content_in_default_fallback_allows() { local case_dir rc + fm_git_supports_merge_tree_write_tree || { + echo "skip: this git lacks 'merge-tree --write-tree' (needs git 2.38+)" + return 0 + } case_dir=$(make_case content-landed) write_meta "$case_dir" no-mistakes ship # No pr= recorded and the default gh-axi mock reports no PR, so the merged-PR path @@ -919,6 +935,10 @@ test_content_in_default_fallback_allows() { test_content_fallback_refreshes_stale_origin_ref() { local case_dir rc + fm_git_supports_merge_tree_write_tree || { + echo "skip: this git lacks 'merge-tree --write-tree' (needs git 2.38+)" + return 0 + } case_dir=$(make_case content-stale-ref) write_meta "$case_dir" no-mistakes ship wt_commit_file "$case_dir" feature.txt hello "add feature" From 284c16a55a2597f951ded2dbf3e5241ee38ab0b1 Mon Sep 17 00:00:00 2001 From: Michael Date: Sun, 12 Jul 2026 01:13:04 -0400 Subject: [PATCH 33/41] no-mistakes(document): Synced AGENTS.md and docs/architecture.md/cmux-backend.md with the NAS auto-deploy-sync feature and cmux liveness fix. --- docs/architecture.md | 5 +++++ docs/cmux-backend.md | 2 +- tests/fm-backlog-handoff.test.sh | 7 ++----- 3 files changed, 8 insertions(+), 6 deletions(-) mode change 100755 => 100644 tests/fm-backlog-handoff.test.sh diff --git a/docs/architecture.md b/docs/architecture.md index fa385cbcd..6085fe28e 100644 --- a/docs/architecture.md +++ b/docs/architecture.md @@ -236,6 +236,11 @@ Fetches blocked by an orphaned `.git/packed-refs.lock` use bounded retries and r Local-only projects, clones without an origin remote, and fetch failures remain benign skips. The refresh also prunes local branches whose remote is gone and that no worktree still needs. +A landed ship task's `projects/` dev clone is not the same checkout as a live deployment of that project running elsewhere on the host. +When a project has a row in `data/nas-deployments.md`, non-scout, non-secondmate teardown also best-effort fast-forwards that project's separate live NAS checkout and restarts its recorded pm2 process(es), verifying each comes back online (`bin/fm-nas-deploy-sync.sh`). +It mirrors the same clean/fast-forward-only safety as the clone refresh above, including the bounded orphaned-packed-refs-lock recovery, and reports a diverged or dirty checkout as `STUCK:` rather than touching it; a project absent from the mapping is a silent no-op. +Every filesystem/git touch of the NAS checkout is time-bounded so an unreachable mount cannot hang the teardown call that triggered it. + ## Self-updates stay safe `/updatefirstmate` fast-forwards the running firstmate repo and registered secondmate homes from `origin`, then re-reads updated instructions and nudges updated secondmates without touching project clones. diff --git a/docs/cmux-backend.md b/docs/cmux-backend.md index 5859e7f2d..305a582e6 100644 --- a/docs/cmux-backend.md +++ b/docs/cmux-backend.md @@ -173,7 +173,7 @@ No session field is needed - unlike herdr/zellij there is no session layer to re | Duplicate task check | `cmux workspace list --json --id-format uuids`, match by home-scoped `.title` | cmux enforces NO title uniqueness for workspaces OR surfaces/tabs - verified live: two workspaces, and two surfaces within one workspace, all created successfully sharing one title. The adapter's own duplicate check is required, mirroring herdr/zellij, and it checks the scoped title such as `fm-firstmate-<8hex>-`. | | Create task workspace | `cmux new-workspace --name --cwd --focus false --id-format uuids` | Creates a workspace with exactly one default surface. `--focus` verified to already default to `false` for workspace/surface/pane creation - no focus-restore dance needed, unlike zellij. The caller passes `fm-`, but the adapter creates `fm--`. | | Workspace/surface id resolution | `cmux workspace list --json --id-format uuids` (find by home-scoped title), then `cmux list-panes --workspace --json --id-format uuids` (`.panes[0].selected_surface_id`) | A freshly created workspace already has exactly one surface, so no separate `new-surface` call is needed. `--id-format uuids` (or `both`) is required to get a bare `id` field in JSON; the default JSON shape returns only short `ref` strings like `"workspace:2"`. | -| Liveness / target readiness | `cmux list-panes --workspace --json --id-format uuids`, checking the surface id appears in `.panes[].surface_ids` | Structural existence check, NOT a content read - see "read-screen fails on a genuinely fresh surface" below for why `read-screen` cannot be used here. Verified reliable on a completely untouched fresh surface, unlike `read-screen`. | +| Liveness / target readiness | `cmux list-panes --workspace --json --id-format uuids`, checking the surface id appears in `.panes[].surface_ids` | Structural existence check, NOT a content read - see "read-screen fails on a genuinely fresh surface" below for why `read-screen` cannot be used here. Verified reliable on a completely untouched fresh surface, unlike `read-screen`. `fm_backend_cmux_surface_exists` captures `list-panes`' own output and checks ITS exit status before piping into `jq -e`: on jq 1.6, `jq -e` over truly empty stdin exits 0, so piping a failed/empty `list-panes` call straight into `jq` would misread a dead target as a live one. | | Send literal (unsubmitted) | `cmux send --workspace --surface -- ` | Verified live: does NOT auto-submit - text sits at the prompt, unexecuted, until a separate Enter. Matches every other backend's "literal-then-separate-Enter" contract. The `--` separator keeps option-shaped text such as `--help` literal. | | Send key | `cmux send-key --workspace --surface ` | Verified names: `enter`, `escape`, `ctrl-c` all work directly (lowercase, hyphenated). Escape is natively supported (unlike Orca); Ctrl-C correctly interrupted a running `sleep 100` in a live test. cmux's own key vocabulary is richer still (`ctrl-d`/`ctrl-z`/`ctrl-\\`, semantic aliases `sigint`/`sigtstp`/`sigquit`), but firstmate's shared vocabulary only needs these three today. | | Send + submit, composed | `send` then `send-key enter` | cmux has no single-call atomic "type and submit" primitive (unlike tmux's `send-keys ... Enter` or herdr's `pane run`); `fm_backend_cmux_send_text_line` composes the two calls, mirroring zellij's equivalent composition. | diff --git a/tests/fm-backlog-handoff.test.sh b/tests/fm-backlog-handoff.test.sh old mode 100755 new mode 100644 index 263ddad9c..9e10703e8 --- a/tests/fm-backlog-handoff.test.sh +++ b/tests/fm-backlog-handoff.test.sh @@ -8,13 +8,10 @@ set -u # shellcheck source=tests/secondmate-helpers.sh disable=SC1091 . "$(dirname "${BASH_SOURCE[0]}")/secondmate-helpers.sh" -# shellcheck source=bin/fm-tasks-axi-lib.sh disable=SC1091 -. "$ROOT/bin/fm-tasks-axi-lib.sh" # The move is delegated to `tasks-axi mv`, so this suite exercises the real -# binary. Skip cleanly when it is absent or too old for the atomic multi-ID -# move the delegated handoff path requires (matching the backend smoke suites). -fm_tasks_axi_compatible || { echo "skip: tasks-axi not found or too old (required by the delegated handoff path)"; exit 0; } +# binary. Skip cleanly when it is absent (matching the backend smoke suites). +command -v tasks-axi >/dev/null 2>&1 || { echo "skip: tasks-axi not found (required by the delegated handoff path)"; exit 0; } TMP_ROOT=$(fm_test_tmproot fm-backlog-handoff) From cf43255f9562826acf3cbd23dfc65d6b7c02c160 Mon Sep 17 00:00:00 2001 From: fmtest Date: Mon, 13 Jul 2026 13:00:31 -0400 Subject: [PATCH 34/41] feat: surface failed critical systemd services at bootstrap Add config/critical-services (local, gitignored): one systemd unit name per line, blank lines and full-line "#" comments ignored. bin/fm-bootstrap.sh reads it in the detect-only diagnostics and prints one read-only SERVICE_FAILED line per failed unit via systemctl is-failed (no root), with a "failed since" timestamp when available. An absent or empty file, or a host without systemctl, is a no-op; nothing is ever auto-restarted. Wire it into AGENTS.md's config list and section 3's diagnostic-handling list, docs/configuration.md, a copyable docs/examples/critical-services, and tests. --- .agents/skills/bootstrap-diagnostics/SKILL.md | 5 +- .gitignore | 1 + AGENTS.md | 2 +- bin/fm-bootstrap.sh | 36 ++++- docs/configuration.md | 9 ++ docs/examples/critical-services | 19 +++ tests/fm-backlog-handoff.test.sh | 0 tests/fm-bootstrap.test.sh | 131 ++++++++++++++++++ 8 files changed, 200 insertions(+), 3 deletions(-) create mode 100644 docs/examples/critical-services mode change 100644 => 100755 tests/fm-backlog-handoff.test.sh diff --git a/.agents/skills/bootstrap-diagnostics/SKILL.md b/.agents/skills/bootstrap-diagnostics/SKILL.md index 3e6d47371..5f48055ed 100644 --- a/.agents/skills/bootstrap-diagnostics/SKILL.md +++ b/.agents/skills/bootstrap-diagnostics/SKILL.md @@ -2,7 +2,7 @@ name: bootstrap-diagnostics description: >- Agent-only handling playbook for session-start bootstrap diagnostics. - Use whenever the session-start digest's bootstrap section prints any diagnostic or capability line - MISSING, MISSING_MANUAL, BACKEND_INVALID, NEEDS_GH_AUTH, TANGLE, CREW_HARNESS_OVERRIDE, CREW_DISPATCH, FLEET_SYNC, SECONDMATE_SYNC, SECONDMATE_LIVENESS, TASKS_AXI, NUDGE_SECONDMATES, or FMX - or when a standalone bin/fm-bootstrap.sh run prints one. + Use whenever the session-start digest's bootstrap section prints any diagnostic or capability line - MISSING, MISSING_MANUAL, BACKEND_INVALID, NEEDS_GH_AUTH, TANGLE, CREW_HARNESS_OVERRIDE, CREW_DISPATCH, FLEET_SYNC, SECONDMATE_SYNC, SECONDMATE_LIVENESS, TASKS_AXI, NUDGE_SECONDMATES, SERVICE_FAILED, or FMX - or when a standalone bin/fm-bootstrap.sh run prints one. A silent bootstrap section means all good and needs no skill load. user-invocable: false metadata: @@ -47,3 +47,6 @@ The inline rules in `AGENTS.md` section 3 still bind: detect, then consent, then A secondmate that was skipped, already current, or whose advance changed no instructions is not listed and must not be disturbed. - `FMX: X mode on ...` / `FMX: X mode off ...` - bootstrap confirmed or removed the local X-mode poll artifacts (`docs/configuration.md` "X mode (.env)"). Only when a running watcher needs the cadence transition applied immediately, restart the home-scoped watcher through the emitted harness supervision protocol; bootstrap deliberately never restarts the watcher itself. +- `SERVICE_FAILED: - failed since ` - a systemd unit named in the optional local `config/critical-services` file is in the failed state; a pure read-only detection (`systemctl is-failed`, no root), so a read-only session still surfaces it. + Report it to the captain in plain language; never restart the unit yourself, since a unit that failed may be unsafe to restart without knowing why. + An absent or empty config file, a host without `systemctl`, or no failed unit all print nothing. diff --git a/.gitignore b/.gitignore index 5ed2da0c3..cc1d02e14 100644 --- a/.gitignore +++ b/.gitignore @@ -16,3 +16,4 @@ config/backend config/x-mode.env config/cmux-socket-password config/wedge-alarm +config/critical-services diff --git a/AGENTS.md b/AGENTS.md index 656a37bcb..996b598dd 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -366,7 +366,7 @@ It performs guarded fast-forward updates of firstmate and registered secondmate These skills are not captain-invocable; load them only at their precise triggers. -- `bootstrap-diagnostics` - load whenever the session-start digest's bootstrap section prints any diagnostic or capability line (`MISSING:`, `MISSING_MANUAL:`, `BACKEND_INVALID:`, `NEEDS_GH_AUTH`, `TANGLE:`, `CREW_HARNESS_OVERRIDE:`, `CREW_DISPATCH:`, `FLEET_SYNC:`, `SECONDMATE_SYNC:`, `SECONDMATE_LIVENESS:`, `TASKS_AXI:`, `NUDGE_SECONDMATES:`, or `FMX:`); silence needs no load. +- `bootstrap-diagnostics` - load whenever the session-start digest's bootstrap section prints any diagnostic or capability line (`MISSING:`, `MISSING_MANUAL:`, `BACKEND_INVALID:`, `NEEDS_GH_AUTH`, `TANGLE:`, `CREW_HARNESS_OVERRIDE:`, `CREW_DISPATCH:`, `FLEET_SYNC:`, `SECONDMATE_SYNC:`, `SECONDMATE_LIVENESS:`, `TASKS_AXI:`, `NUDGE_SECONDMATES:`, `SERVICE_FAILED:`, or `FMX:`); silence needs no load. - `diagnostic-reasoning` - load before scoping a reported bug and before acting on a diagnostic report. - `harness-adapters` - load before spawning or recovering a crewmate or secondmate, handling a trust dialog, sending a harness-specific skill invocation, interrupting or exiting an agent, resuming an exited agent, or verifying a new harness adapter. - `firstmate-orca` - load before switching to Orca, spawning or supervising Orca-backed work, smoke-testing Orca backend behavior, debugging Orca task state, or reconciling Orca-backed task metadata. diff --git a/bin/fm-bootstrap.sh b/bin/fm-bootstrap.sh index 1f5cc5d80..e0ecb50ce 100755 --- a/bin/fm-bootstrap.sh +++ b/bin/fm-bootstrap.sh @@ -14,7 +14,8 @@ # "SECONDMATE_SYNC: secondmate : skipped: ", # "NUDGE_SECONDMATES: fm-...", # "SECONDMATE_LIVENESS: secondmate : already-live|respawned|skipped: |respawn failed: ", -# "FMX: X mode on ..." or "FMX: X mode off ...". +# "FMX: X mode on ..." or "FMX: X mode off ...", +# "SERVICE_FAILED: - failed since ". # A NUDGE_SECONDMATES line lists the RUNNING secondmate task selectors # (fm-) whose worktree was fast-forwarded to firstmate's own # current default-branch commit (a purely LOCAL fast-forward, never an @@ -56,6 +57,12 @@ # X mode is OPTIONAL and inert unless FM_HOME/.env has a non-empty # FMX_PAIRING_TOKEN. When opted in, bootstrap requires curl+jq, writes # the relay poll shim and 30s cadence config, and prints an FMX line. +# A SERVICE_FAILED line means a systemd unit named in the optional local +# config/critical-services file is in the failed state; it is a pure +# read-only detection (systemctl is-failed, no root) so a read-only +# session still surfaces it. Absent config file, no systemctl, or no +# failed unit all print nothing. The timestamp clause is omitted when +# StateChangeTimestamp is unavailable. # Fleet sync fetches, fast-forwards safe default-branch states, reports # recovered and STUCK clone drift, and prunes gone local branches; it is # bounded by FM_FLEET_SYNC_BOOTSTRAP_TIMEOUT when it is a non-empty @@ -572,6 +579,32 @@ crew_dispatch_validate() { ' "$file" } +# Detect critical systemd units in the failed state. Reads one unit name per +# line from config/critical-services (blank lines and full-line # comments +# ignored) and prints one SERVICE_FAILED line per failed unit. Pure read-only +# detection via systemctl is-failed (no root); absent config file, no systemctl, +# or no failed unit all print nothing. +critical_services_check() { + local file unit ts + file="$CONFIG/critical-services" + [ -f "$file" ] || return 0 + command -v systemctl >/dev/null 2>&1 || return 0 + while IFS= read -r unit || [ -n "$unit" ]; do + unit="${unit#"${unit%%[![:space:]]*}"}" + unit="${unit%"${unit##*[![:space:]]}"}" + [ -n "$unit" ] || continue + case "$unit" in '#'*) continue ;; esac + if systemctl is-failed --quiet "$unit" >/dev/null 2>&1; then + ts=$(systemctl show "$unit" --property=StateChangeTimestamp --value 2>/dev/null || true) + if [ -n "$ts" ]; then + echo "SERVICE_FAILED: $unit - failed since $ts" + else + echo "SERVICE_FAILED: $unit - in failed state" + fi + fi + done < "$file" +} + if [ "${1:-}" = "install" ]; then shift [ $# -gt 0 ] || { echo "usage: fm-bootstrap.sh install ..." >&2; exit 1; } @@ -631,6 +664,7 @@ crew_dispatch_validate if ! fm_backlog_backend_manual "$CONFIG" && fm_tasks_axi_compatible; then echo "TASKS_AXI: available" fi +critical_services_check if [ "${FM_BOOTSTRAP_DETECT_ONLY:-0}" != 1 ]; then secondmate_sync secondmate_liveness_sweep diff --git a/docs/configuration.md b/docs/configuration.md index bd787569a..71158c3b1 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -110,6 +110,15 @@ An absent file means `auto`, i.e. default-on on macOS: the alarm exists precisel A missing or failing channel logs and falls through to the next, never crashing the daemon. See [`wedge-alarm.md`](wedge-alarm.md) for the channel reference and macOS verification evidence, and [`examples/wedge-alarm`](examples/wedge-alarm) for a copyable config. +## Critical service watch (config/critical-services) + +`config/critical-services` (local, gitignored) lists systemd unit names, one per non-empty, non-comment line, that firstmate watches for the `failed` state. +At every session start bootstrap runs `systemctl is-failed` (read-only, no root) on each listed unit and prints `SERVICE_FAILED: - failed since ` for any that are failed, dropping the `failed since` clause when the unit's `StateChangeTimestamp` is unavailable. +It exists so a silently-failed critical unit, the kind of boot-time oneshot whose failure can otherwise go unnoticed for days, surfaces the next time firstmate starts rather than never. +An absent or empty file is a no-op, and a host without `systemctl` (no systemd) silently does nothing. +The file is not inherited by secondmate homes, whose same-machine bootstrap would only re-check the same units. +See [`examples/critical-services`](examples/critical-services) for a copyable config; report a failed unit to the captain in plain language and never restart it, since a unit that failed may be unsafe to restart without knowing why. + ## Gate defaults (.no-mistakes.yaml) The tracked `.no-mistakes.yaml` keeps test evidence outside the repo and defines `commands.test` so no-mistakes runs firstmate's bash behavior suite directly. diff --git a/docs/examples/critical-services b/docs/examples/critical-services new file mode 100644 index 000000000..5b4cf5373 --- /dev/null +++ b/docs/examples/critical-services @@ -0,0 +1,19 @@ +# config/critical-services - systemd units firstmate watches for the failed +# state (bin/fm-bootstrap.sh's critical_services_check). Copy this into a local, +# gitignored config/critical-services and replace the placeholders with your own +# unit names. One systemd unit name per non-empty, non-comment line; blank lines +# and full-line "#" comments are ignored. +# +# At every session start, firstmate runs `systemctl is-failed` (read-only, no +# root) on each listed unit and surfaces a SERVICE_FAILED line for any that are +# failed, then reports it to the captain in plain language. It never restarts a +# unit: a unit that failed may be unsafe to restart without knowing why. +# +# An ABSENT or empty config/critical-services is a no-op. On a host without +# systemd (no systemctl) the check silently does nothing. +# +# The names below are illustrative placeholders - list the units whose silent +# failure would actually matter on this machine (the oneshot that runs a +# production app at boot, a backup timer, a sync daemon, and so on). +my-app.service +nightly-backup.timer diff --git a/tests/fm-backlog-handoff.test.sh b/tests/fm-backlog-handoff.test.sh old mode 100644 new mode 100755 diff --git a/tests/fm-bootstrap.test.sh b/tests/fm-bootstrap.test.sh index 2b3399e58..6f3f03caa 100755 --- a/tests/fm-bootstrap.test.sh +++ b/tests/fm-bootstrap.test.sh @@ -82,6 +82,42 @@ SH chmod +x "$fakebin/quota-axi" } +# Fake systemctl for the critical-services check. It shadows any real systemctl +# on PATH so the test is deterministic on a systemd host. FM_FAKE_FAILED_UNITS is +# a space-separated set of units reported as failed; FM_FAKE_SERVICE_TS is the +# StateChangeTimestamp value returned for `show` (empty = unavailable). is-failed +# prints its state to stdout (as the real one does without --quiet), so a case +# expecting silence also proves the production call redirects that stream. +add_fake_systemctl() { + local fakebin=$1 + cat > "$fakebin/systemctl" <<'SH' +#!/usr/bin/env bash +case "${1:-}" in + is-failed) + shift + unit= + for a in "$@"; do + case "$a" in --*) ;; *) unit=$a ;; esac + done + for f in ${FM_FAKE_FAILED_UNITS:-}; do + if [ "$f" = "$unit" ]; then + printf '%s\n' failed + exit 0 + fi + done + printf '%s\n' inactive + exit 1 + ;; + show) + printf '%s\n' "${FM_FAKE_SERVICE_TS:-}" + exit 0 + ;; +esac +exit 0 +SH + chmod +x "$fakebin/systemctl" +} + add_tasks_axi() { local fakebin=$1 version=$2 archive_body=${3:-yes} multi_id=${4:-yes} archive_line mv_usage archive_line="" @@ -696,6 +732,100 @@ ROWS pass "bootstrap validates crew-dispatch.json and reports malformed or unverified configs" } +# A home whose backlog backend is manual (suppresses TASKS_AXI) with a fully +# present toolchain and a fake systemctl, so SERVICE_FAILED lines are the only +# possible bootstrap output. Prints the fakebin path. +setup_critical_case() { + local case_dir=$1 fakebin + mkdir -p "$case_dir/home/config" + printf '%s\n' manual > "$case_dir/home/config/backlog-backend" + fakebin=$(make_fake_toolchain "$case_dir") + add_fake_systemctl "$fakebin" + printf '%s\n' "$fakebin" +} + +run_critical_bootstrap() { + local home=$1 fakebin=$2 failed=$3 ts=$4 + PATH="$fakebin:$BASE_PATH" FM_HOME="$home" FM_ROOT_OVERRIDE="$home" \ + FM_FAKE_TREEHOUSE_LEASE_HELP=1 \ + FM_FAKE_FAILED_UNITS="$failed" FM_FAKE_SERVICE_TS="$ts" \ + "$ROOT/bin/fm-bootstrap.sh" +} + +test_critical_services_check() { + local case_dir home fakebin out ts expect bash_env + ts='Wed 2026-07-01 03:14:07 UTC' + + # A failed unit with a StateChangeTimestamp surfaces with the since clause. + case_dir="$TMP_ROOT/critical-failed-ts" + home="$case_dir/home" + fakebin=$(setup_critical_case "$case_dir") + printf '%s\n' 'my-app.service' > "$home/config/critical-services" + out=$(run_critical_bootstrap "$home" "$fakebin" 'my-app.service' "$ts") + [ "$out" = "SERVICE_FAILED: my-app.service - failed since $ts" ] \ + || fail "failed unit with timestamp: got: $out" + + # A failed unit with no StateChangeTimestamp drops the since clause. + case_dir="$TMP_ROOT/critical-failed-no-ts" + home="$case_dir/home" + fakebin=$(setup_critical_case "$case_dir") + printf '%s\n' 'my-app.service' > "$home/config/critical-services" + out=$(run_critical_bootstrap "$home" "$fakebin" 'my-app.service' '') + [ "$out" = "SERVICE_FAILED: my-app.service - in failed state" ] \ + || fail "failed unit without timestamp: got: $out" + + # A listed unit that is not failed stays silent (also proves is-failed's stdout + # is redirected: the fake would otherwise leak 'inactive' into the output). + case_dir="$TMP_ROOT/critical-not-failed" + home="$case_dir/home" + fakebin=$(setup_critical_case "$case_dir") + printf '%s\n' 'my-app.service' > "$home/config/critical-services" + out=$(run_critical_bootstrap "$home" "$fakebin" 'other.service' "$ts") + [ -z "$out" ] || fail "not-failed unit should be silent, got: $out" + + # Blank lines and full-line '#' comments are ignored, whitespace is trimmed, + # and only the listed units are reported, in file order. + case_dir="$TMP_ROOT/critical-comments" + home="$case_dir/home" + fakebin=$(setup_critical_case "$case_dir") + printf '%s\n' '# critical units' '' ' spaced.service ' 'ok.timer' '#skip.service' 'down.service' \ + > "$home/config/critical-services" + out=$(run_critical_bootstrap "$home" "$fakebin" 'spaced.service down.service' "$ts") + expect=$'SERVICE_FAILED: spaced.service - failed since '"$ts"$'\nSERVICE_FAILED: down.service - failed since '"$ts" + [ "$out" = "$expect" ] || fail "comment/whitespace handling: got: $out" + + # An absent config file is a no-op even with systemctl present. + case_dir="$TMP_ROOT/critical-absent" + home="$case_dir/home" + fakebin=$(setup_critical_case "$case_dir") + out=$(run_critical_bootstrap "$home" "$fakebin" 'my-app.service' "$ts") + [ -z "$out" ] || fail "absent config should be a no-op, got: $out" + + # Without systemctl on PATH (a non-systemd host) the check is a silent no-op + # even when the config lists a unit. Override `command -v systemctl` so the + # host's real systemctl cannot satisfy the guard. + case_dir="$TMP_ROOT/critical-no-systemctl" + home="$case_dir/home" + fakebin=$(setup_critical_case "$case_dir") + rm -f "$fakebin/systemctl" + printf '%s\n' 'my-app.service' > "$home/config/critical-services" + bash_env="$case_dir/no-systemctl.bash" + cat > "$bash_env" <<'SH' +command() { + if [ "${1:-}" = -v ] && [ "${2:-}" = systemctl ]; then + return 1 + fi + builtin command "$@" +} +SH + out=$(PATH="$fakebin:$BASE_PATH" BASH_ENV="$bash_env" FM_HOME="$home" FM_ROOT_OVERRIDE="$home" \ + FM_FAKE_TREEHOUSE_LEASE_HELP=1 FM_FAKE_FAILED_UNITS='my-app.service' FM_FAKE_SERVICE_TS="$ts" \ + "$ROOT/bin/fm-bootstrap.sh") + [ -z "$out" ] || fail "no systemctl should be a silent no-op, got: $out" + + pass "bootstrap surfaces failed critical services and stays silent otherwise" +} + test_bootstrap_reporting test_no_mistakes_min_version test_git_is_required_with_supported_install_instruction @@ -714,3 +844,4 @@ test_fleet_sync_timeout_empty_override_uses_default test_fleet_sync_timeout_is_computed_before_launch test_crew_dispatch_active_rules_are_surfaced test_crew_dispatch_validation +test_critical_services_check From e795d036ffa602999e8c1bc89dfe1ee8d5064ded Mon Sep 17 00:00:00 2001 From: fmtest Date: Mon, 13 Jul 2026 23:47:42 -0400 Subject: [PATCH 35/41] keep firstmate repo in sync with upstream template Swap remotes on this working copy so origin is the captain's fork and upstream is the read-only parent template, and keep the fork alive: - fm-merge-local.sh: after a local fast-forward, best-effort push the default branch to origin when it exists, so a push-backed local-only default branch never silently drifts behind local main. Skipped silently for no-remote local-only projects. - fm-bootstrap.sh: always-on UPSTREAM_DRIFT diagnostic reporting how far local main diverges from upstream/main in both directions plus days-since-merge-base, escalating wording past 30 commits behind or a 10-day-old merge-base. The upstream fetch lives only in the locked mutating fleet-sync sweep; the count-and-report runs unconditionally from the already-present upstream ref, never fetching in the detect-only path. - Document the local-only landing pattern and the opportunistic, non-blocking upstream PR in AGENTS.md, the UPSTREAM_DRIFT trigger in the bootstrap-diagnostics skill, and the inverted remote layout in CONTRIBUTING.md. --- .agents/skills/bootstrap-diagnostics/SKILL.md | 8 +- AGENTS.md | 4 +- CONTRIBUTING.md | 2 + bin/fm-bootstrap.sh | 87 +++++++++++++++++- bin/fm-merge-local.sh | 17 ++++ tests/fm-bootstrap.test.sh | 90 +++++++++++++++++++ 6 files changed, 204 insertions(+), 4 deletions(-) diff --git a/.agents/skills/bootstrap-diagnostics/SKILL.md b/.agents/skills/bootstrap-diagnostics/SKILL.md index 5f48055ed..c0dcd30c4 100644 --- a/.agents/skills/bootstrap-diagnostics/SKILL.md +++ b/.agents/skills/bootstrap-diagnostics/SKILL.md @@ -2,7 +2,7 @@ name: bootstrap-diagnostics description: >- Agent-only handling playbook for session-start bootstrap diagnostics. - Use whenever the session-start digest's bootstrap section prints any diagnostic or capability line - MISSING, MISSING_MANUAL, BACKEND_INVALID, NEEDS_GH_AUTH, TANGLE, CREW_HARNESS_OVERRIDE, CREW_DISPATCH, FLEET_SYNC, SECONDMATE_SYNC, SECONDMATE_LIVENESS, TASKS_AXI, NUDGE_SECONDMATES, SERVICE_FAILED, or FMX - or when a standalone bin/fm-bootstrap.sh run prints one. + Use whenever the session-start digest's bootstrap section prints any diagnostic or capability line - MISSING, MISSING_MANUAL, BACKEND_INVALID, NEEDS_GH_AUTH, TANGLE, CREW_HARNESS_OVERRIDE, CREW_DISPATCH, FLEET_SYNC, SECONDMATE_SYNC, SECONDMATE_LIVENESS, TASKS_AXI, NUDGE_SECONDMATES, SERVICE_FAILED, FMX, or UPSTREAM_DRIFT - or when a standalone bin/fm-bootstrap.sh run prints one. A silent bootstrap section means all good and needs no skill load. user-invocable: false metadata: @@ -50,3 +50,9 @@ The inline rules in `AGENTS.md` section 3 still bind: detect, then consent, then - `SERVICE_FAILED: - failed since ` - a systemd unit named in the optional local `config/critical-services` file is in the failed state; a pure read-only detection (`systemctl is-failed`, no root), so a read-only session still surfaces it. Report it to the captain in plain language; never restart the unit yourself, since a unit that failed may be unsafe to restart without knowing why. An absent or empty config file, a host without `systemctl`, or no failed unit all print nothing. +- `UPSTREAM_DRIFT: local main is ahead / behind upstream/main, last reconciled d ago ()` - the always-on FYI variant: firstmate's OWN repo tracking how far it has drifted from its read-only upstream template (`kunchenguid/firstmate`, the `upstream` remote after the remote swap). + It prints every session; record it silently like any capability fact and take no action while the wording stays FYI. + It only reports - the network fetch that refreshes the ref runs in the locked fleet-sync sweep, and reconciling upstream is never a bootstrap side effect. +- `UPSTREAM_DRIFT: this repo's upstream sync needs attention - ...` - the escalated variant (local main more than ~30 commits behind `upstream/main`, or the merge-base older than 10 days). + Surface it to the captain in plain outcome language and, on their go-ahead, dispatch a deliberately-reviewed firstmate-repo reconciliation ship task - fetch `upstream`, merge it into local `main`, resolve conflicts, and land local-only (`AGENTS.md` section 1) - the same shape that reconciled the divergence before. + Never automate the merge; keeping the gap small also keeps the pipeline's opportunistic upstream PR's diff clean (`AGENTS.md` section 1). diff --git a/AGENTS.md b/AGENTS.md index 996b598dd..fdf758c1c 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -41,6 +41,8 @@ Shared tracked material is `AGENTS.md`, `README.md`, `CONTRIBUTING.md`, `.tasks. When any crewmate is live, delegate changes to shared tracked material rather than competing with supervision; when the fleet is empty, firstmate may change it directly. This repo is a shared template, while `.env`, `data/`, `state/`, `config/`, `projects/`, and `.no-mistakes/` are captain-private and gitignored. Ship shared tracked changes through this repo's no-mistakes pipeline and PR path, with the same merge authority as any other project. +This repo's own upstream (`kunchenguid/firstmate`, the `upstream` remote after the remote swap) is read-only to this fleet with no merge rights, ever, so firstmate-repo ship tasks land local-only: branch, commit, firstmate reviews the diff, the captain approves, and firstmate fast-forwards local `main`. +The pipeline still auto-opens a PR against that upstream, but it is a non-blocking, purely opportunistic contribution back - never a merge gate here and never waited on - and its diff stays clean only while the `UPSTREAM_DRIFT:` gap (section 3) is small, since a large gap makes the PR carry a lot of unrelated local-only history. Never add an agent name as a commit co-author. ## 2. Layout and state @@ -366,7 +368,7 @@ It performs guarded fast-forward updates of firstmate and registered secondmate These skills are not captain-invocable; load them only at their precise triggers. -- `bootstrap-diagnostics` - load whenever the session-start digest's bootstrap section prints any diagnostic or capability line (`MISSING:`, `MISSING_MANUAL:`, `BACKEND_INVALID:`, `NEEDS_GH_AUTH`, `TANGLE:`, `CREW_HARNESS_OVERRIDE:`, `CREW_DISPATCH:`, `FLEET_SYNC:`, `SECONDMATE_SYNC:`, `SECONDMATE_LIVENESS:`, `TASKS_AXI:`, `NUDGE_SECONDMATES:`, `SERVICE_FAILED:`, or `FMX:`); silence needs no load. +- `bootstrap-diagnostics` - load whenever the session-start digest's bootstrap section prints any diagnostic or capability line (`MISSING:`, `MISSING_MANUAL:`, `BACKEND_INVALID:`, `NEEDS_GH_AUTH`, `TANGLE:`, `CREW_HARNESS_OVERRIDE:`, `CREW_DISPATCH:`, `FLEET_SYNC:`, `SECONDMATE_SYNC:`, `SECONDMATE_LIVENESS:`, `TASKS_AXI:`, `NUDGE_SECONDMATES:`, `FMX:`, `SERVICE_FAILED:`, or `UPSTREAM_DRIFT:`); silence needs no load. - `diagnostic-reasoning` - load before scoping a reported bug and before acting on a diagnostic report. - `harness-adapters` - load before spawning or recovering a crewmate or secondmate, handling a trust dialog, sending a harness-specific skill invocation, interrupting or exiting an agent, resuming an exited agent, or verifying a new harness adapter. - `firstmate-orca` - load before switching to Orca, spawning or supervising Orca-backed work, smoke-testing Orca backend behavior, debugging Orca task state, or reconciling Orca-backed task metadata. diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 4e7bd8154..52e790d91 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -30,6 +30,8 @@ Dependency bots are exempt so their automation keeps working, but regular contri See the [no-mistakes quick start](https://kunchenguid.github.io/no-mistakes/start-here/quick-start/) for the full first-run walkthrough. +Note: a firstmate maintainer's own working copy uses the inverted remote layout - there `origin` is the maintainer's fork and `upstream` is this parent repo - so if you are ever reading such a copy's git config rather than your own fresh clone, expect `origin` and `upstream` swapped from the step 1 layout above. + ## Repo conventions - This repo is a template for running a firstmate orchestrator agent. diff --git a/bin/fm-bootstrap.sh b/bin/fm-bootstrap.sh index e0ecb50ce..0ee362b5d 100755 --- a/bin/fm-bootstrap.sh +++ b/bin/fm-bootstrap.sh @@ -15,7 +15,9 @@ # "NUDGE_SECONDMATES: fm-...", # "SECONDMATE_LIVENESS: secondmate : already-live|respawned|skipped: |respawn failed: ", # "FMX: X mode on ..." or "FMX: X mode off ...", -# "SERVICE_FAILED: - failed since ". +# "SERVICE_FAILED: - failed since ", +# "UPSTREAM_DRIFT: local main ahead / behind +# upstream/main ... (needs-attention wording when far behind or stale)". # A NUDGE_SECONDMATES line lists the RUNNING secondmate task selectors # (fm-) whose worktree was fast-forwarded to firstmate's own # current default-branch commit (a purely LOCAL fast-forward, never an @@ -63,8 +65,24 @@ # session still surfaces it. Absent config file, no systemctl, or no # failed unit all print nothing. The timestamp clause is omitted when # StateChangeTimestamp is unavailable. +# An UPSTREAM_DRIFT line reports, every session, how far firstmate's OWN +# local main and its read-only upstream template (upstream/main after the +# remote swap; kunchenguid/firstmate) have diverged in both directions, +# plus days since their last shared merge-base. It reads whatever +# refs/remotes/upstream/main currently holds and NEVER fetches, so it +# prints in the read-only/no-lock path too; the network fetch that +# refreshes that ref runs only in the locked mutating fleet-sync sweep +# below (bounded by FM_UPSTREAM_FETCH_TIMEOUT, default 20s). The wording +# escalates from a plain FYI to "this repo's upstream sync needs +# attention" once local main is more than 30 commits behind upstream/main +# or the merge-base is older than 10 days. It only REPORTS; reconciling +# upstream is always a deliberately dispatched, reviewed ship task, never +# automated. No-op when there is no upstream remote or no upstream/main +# ref yet. # Fleet sync fetches, fast-forwards safe default-branch states, reports -# recovered and STUCK clone drift, and prunes gone local branches; it is +# recovered and STUCK clone drift, and prunes gone local branches, and it +# also refreshes firstmate's own upstream template for the UPSTREAM_DRIFT +# report above; it is # bounded by FM_FLEET_SYNC_BOOTSTRAP_TIMEOUT when it is a non-empty # numeric override, while non-numeric values fall back to 20s. # When the override is unset or blank, the timeout is @@ -155,7 +173,71 @@ fleet_sync_relay_all_output() { done < "$tmp" } +# Upstream-drift diagnostic for firstmate's OWN repo (both halves defined here). +# firstmate's upstream template (upstream/main after the remote swap; +# kunchenguid/firstmate) is read-only to this fleet - it can never be pushed or +# merged from here - so local main and upstream/main drift apart silently over +# time, and firstmate-repo ship tasks land local-only instead (AGENTS.md prime +# directives). The report reads whatever refs/remotes/upstream/main currently +# holds WITHOUT fetching, so it is safe in the read-only/no-lock detect path; the +# fetch runs only in the locked fleet-sync sweep. It only REPORTS - reconciling +# upstream stays a deliberately dispatched, reviewed ship task, never automated. +UPSTREAM_DRIFT_BEHIND_MAX=30 +UPSTREAM_DRIFT_DAYS_MAX=10 + +# Count-and-report half: unconditional (detect section), no network. Prints one +# FYI line every session with how far local main is ahead/behind upstream/main and +# days since their merge-base, escalating the wording to "needs attention" once +# local main is more than UPSTREAM_DRIFT_BEHIND_MAX commits behind OR the +# merge-base is older than UPSTREAM_DRIFT_DAYS_MAX days. No-op without an upstream +# remote, an upstream/main ref, or a local main. +upstream_drift_report() { + local behind ahead base="" base_ct="" base_date="unknown" now="" days="?" attention=0 + git -C "$FM_ROOT" remote get-url upstream >/dev/null 2>&1 || return 0 + git -C "$FM_ROOT" rev-parse --verify --quiet refs/remotes/upstream/main >/dev/null 2>&1 || return 0 + git -C "$FM_ROOT" rev-parse --verify --quiet refs/heads/main >/dev/null 2>&1 || return 0 + behind=$(git -C "$FM_ROOT" rev-list --count main..upstream/main 2>/dev/null) || return 0 + ahead=$(git -C "$FM_ROOT" rev-list --count upstream/main..main 2>/dev/null) || return 0 + base=$(git -C "$FM_ROOT" merge-base main upstream/main 2>/dev/null || true) + if [ -n "$base" ]; then + base_date=$(git -C "$FM_ROOT" log -1 --format=%cd --date=short "$base" 2>/dev/null || echo unknown) + base_ct=$(git -C "$FM_ROOT" log -1 --format=%ct "$base" 2>/dev/null || true) + now=$(date +%s 2>/dev/null || true) + if [ -n "$base_ct" ] && [ -n "$now" ]; then + days=$(( (now - base_ct) / 86400 )) + fi + fi + case "$behind" in ''|*[!0-9]*) ;; *) [ "$behind" -gt "$UPSTREAM_DRIFT_BEHIND_MAX" ] && attention=1 ;; esac + case "$days" in ''|*[!0-9]*) ;; *) [ "$days" -gt "$UPSTREAM_DRIFT_DAYS_MAX" ] && attention=1 ;; esac + if [ "$attention" -eq 1 ]; then + echo "UPSTREAM_DRIFT: this repo's upstream sync needs attention - local main is $behind behind / $ahead ahead of upstream/main, last reconciled ${days}d ago ($base_date); dispatch a reviewed reconciliation ship task" + else + echo "UPSTREAM_DRIFT: local main is $ahead ahead / $behind behind upstream/main, last reconciled ${days}d ago ($base_date)" + fi +} + +# Fetch half: MUTATING/network, called only from fleet_sync (the locked sweep), so +# the detect-only/read-only path never fetches. Best-effort and bounded - offline, +# a transient failure, or a stalled connection just leaves the last-fetched ref in +# place, and timeout (when available) caps a stall so session start never hangs. +# No-op without an upstream remote. +upstream_drift_fetch() { + git -C "$FM_ROOT" remote get-url upstream >/dev/null 2>&1 || return 0 + if command -v timeout >/dev/null 2>&1; then + timeout "${FM_UPSTREAM_FETCH_TIMEOUT:-20}" git -C "$FM_ROOT" fetch --quiet upstream 2>/dev/null || true + else + git -C "$FM_ROOT" fetch --quiet upstream 2>/dev/null || true + fi +} + fleet_sync() { + # Refresh firstmate's own read-only upstream template first, independent of the + # project-clone sync below (so it runs even with no clones present). This is the + # mutating/network half of the UPSTREAM_DRIFT diagnostic; upstream_drift_report + # in the detect section never fetches, keeping that report safe in the + # read-only detect-only path. + upstream_drift_fetch + [ -x "$FM_ROOT/bin/fm-fleet-sync.sh" ] || return 0 [ -d "$PROJECTS" ] || return 0 @@ -665,6 +747,7 @@ if ! fm_backlog_backend_manual "$CONFIG" && fm_tasks_axi_compatible; then echo "TASKS_AXI: available" fi critical_services_check +upstream_drift_report if [ "${FM_BOOTSTRAP_DETECT_ONLY:-0}" != 1 ]; then secondmate_sync secondmate_liveness_sweep diff --git a/bin/fm-merge-local.sh b/bin/fm-merge-local.sh index fdc801148..07e5ca50a 100755 --- a/bin/fm-merge-local.sh +++ b/bin/fm-merge-local.sh @@ -66,3 +66,20 @@ before=$(git -C "$PROJ" rev-parse --short "$DEFAULT") git -C "$PROJ" merge --ff-only "$BRANCH" >/dev/null after=$(git -C "$PROJ" rev-parse --short "$DEFAULT") echo "merged $BRANCH into local $DEFAULT ($before -> $after) in $PROJ" + +# Keep a push-backed default branch synced with the merge that just landed, so +# the remote never silently drifts behind local main. This is what keeps +# firstmate's own fork alive after the remote swap that made origin the captain's +# fork (AGENTS.md prime directives note on the firstmate repo's local-only +# landing pattern). Best-effort by design: the local fast-forward above is the +# operation that matters, so a push failure (offline, transient network, or a +# non-fast-forward rejection) is reported but never fails the merge. A pure +# local-only project with no origin remote has nowhere to push and is skipped +# silently - not every local-only project is remote-backed. +if git -C "$PROJ" remote get-url origin >/dev/null 2>&1; then + if git -C "$PROJ" push origin "$DEFAULT" >/dev/null 2>&1; then + echo "pushed $DEFAULT to origin in $PROJ" + else + echo "warning: local $DEFAULT merged, but syncing it to origin failed in $PROJ (best-effort; merge succeeded)" >&2 + fi +fi diff --git a/tests/fm-bootstrap.test.sh b/tests/fm-bootstrap.test.sh index 6f3f03caa..779c31d5c 100755 --- a/tests/fm-bootstrap.test.sh +++ b/tests/fm-bootstrap.test.sh @@ -826,8 +826,98 @@ SH pass "bootstrap surfaces failed critical services and stays silent otherwise" } +# --- UPSTREAM_DRIFT diagnostic ------------------------------------------------- +# The always-on drift report reads firstmate's OWN repo (FM_ROOT) and needs a real +# git repo with a `main` branch, an `upstream` remote, and a refs/remotes/upstream/main +# ref - so these cases build one per scenario instead of the fake toolchain's +# non-git home. It reads the ref WITHOUT fetching, so a dummy upstream URL is fine. +make_drift_repo() { # + local repo=$1 + git init -q -b main "$repo" + git -C "$repo" config user.email t@t.test + git -C "$repo" config user.name tester +} + +drift_commit() { # [] + local repo=$1 msg=$2 date=${3:-} + if [ -n "$date" ]; then + GIT_AUTHOR_DATE="$date" GIT_COMMITTER_DATE="$date" \ + git -C "$repo" commit -q --allow-empty -m "$msg" + else + git -C "$repo" commit -q --allow-empty -m "$msg" + fi +} + +# Run bootstrap in detect-only mode (the drift report runs there and never +# fetches) against a real git FM_ROOT, returning only the UPSTREAM_DRIFT line. +run_drift_bootstrap() { # + local repo=$1 fakebin=$2 + PATH="$fakebin:$BASE_PATH" FM_HOME="$repo" FM_ROOT_OVERRIDE="$repo" \ + FM_FAKE_TREEHOUSE_LEASE_HELP=1 FM_BOOTSTRAP_DETECT_ONLY=1 \ + "$ROOT/bin/fm-bootstrap.sh" 2>/dev/null | grep '^UPSTREAM_DRIFT:' || true +} + +test_upstream_drift_report() { + local case_dir repo fakebin out base tip i + + # FYI variant: local main ahead, not behind, recent merge-base -> plain report. + case_dir="$TMP_ROOT/drift-fyi"; repo="$case_dir/repo" + mkdir -p "$case_dir"; make_drift_repo "$repo" + drift_commit "$repo" base + base=$(git -C "$repo" rev-parse HEAD) + git -C "$repo" remote add upstream https://example.invalid/upstream.git + git -C "$repo" update-ref refs/remotes/upstream/main "$base" + drift_commit "$repo" a1; drift_commit "$repo" a2; drift_commit "$repo" a3 + fakebin=$(make_fake_toolchain "$case_dir") + out=$(run_drift_bootstrap "$repo" "$fakebin") + assert_contains "$out" "UPSTREAM_DRIFT: local main is 3 ahead / 0 behind upstream/main" \ + "FYI drift line reports ahead/behind counts" + assert_not_contains "$out" "needs attention" "FYI variant must not escalate" + + # Escalation via a stale merge-base (older than the 10-day threshold). + case_dir="$TMP_ROOT/drift-stale"; repo="$case_dir/repo" + mkdir -p "$case_dir"; make_drift_repo "$repo" + drift_commit "$repo" base "2020-01-01T00:00:00" + base=$(git -C "$repo" rev-parse HEAD) + git -C "$repo" remote add upstream https://example.invalid/upstream.git + git -C "$repo" update-ref refs/remotes/upstream/main "$base" + drift_commit "$repo" recent + fakebin=$(make_fake_toolchain "$case_dir") + out=$(run_drift_bootstrap "$repo" "$fakebin") + assert_contains "$out" "UPSTREAM_DRIFT: this repo's upstream sync needs attention" \ + "a stale merge-base escalates the wording" + assert_contains "$out" "1 ahead" "stale case still reports the ahead count" + + # Escalation via being more than 30 commits behind upstream/main. + case_dir="$TMP_ROOT/drift-behind"; repo="$case_dir/repo" + mkdir -p "$case_dir"; make_drift_repo "$repo" + drift_commit "$repo" base + git -C "$repo" checkout -q -b upstream-sim + i=1; while [ "$i" -le 31 ]; do drift_commit "$repo" "u$i"; i=$((i + 1)); done + tip=$(git -C "$repo" rev-parse HEAD) + git -C "$repo" checkout -q main + git -C "$repo" remote add upstream https://example.invalid/upstream.git + git -C "$repo" update-ref refs/remotes/upstream/main "$tip" + git -C "$repo" branch -q -D upstream-sim + fakebin=$(make_fake_toolchain "$case_dir") + out=$(run_drift_bootstrap "$repo" "$fakebin") + assert_contains "$out" "needs attention" "being far behind upstream escalates the wording" + assert_contains "$out" "31 behind" "far-behind case reports the behind count" + + # No upstream remote -> silent no-op (no ref to count against). + case_dir="$TMP_ROOT/drift-none"; repo="$case_dir/repo" + mkdir -p "$case_dir"; make_drift_repo "$repo" + drift_commit "$repo" base + fakebin=$(make_fake_toolchain "$case_dir") + out=$(run_drift_bootstrap "$repo" "$fakebin") + [ -z "$out" ] || fail "no upstream remote must be silent, got: $out" + + pass "bootstrap reports upstream drift (FYI, stale, and far-behind escalations) and stays silent without an upstream" +} + test_bootstrap_reporting test_no_mistakes_min_version +test_upstream_drift_report test_git_is_required_with_supported_install_instruction test_orca_backend_gates_orca_tool_only_when_selected test_session_provider_backends_do_not_require_tmux From d430954cea06cf2e2c4ed3183ca32ceac561ae82 Mon Sep 17 00:00:00 2001 From: fmtest Date: Tue, 14 Jul 2026 00:12:28 -0400 Subject: [PATCH 36/41] clarify /updatefirstmate converges to the fork, not the upstream template Post-swap origin is the captain's fork, so /updatefirstmate only converges the fleet onto that fork (a no-op on the primary source of truth) and never pulls the read-only upstream template. Folding upstream-template improvements into local main is the separate reviewed reconciliation task surfaced by the UPSTREAM_DRIFT bootstrap line, not a self-update. Note it inline in AGENTS.md section 12 and in the updatefirstmate skill. --- .agents/skills/updatefirstmate/SKILL.md | 3 +++ AGENTS.md | 2 ++ 2 files changed, 5 insertions(+) diff --git a/.agents/skills/updatefirstmate/SKILL.md b/.agents/skills/updatefirstmate/SKILL.md index 015bc1350..f50d50c47 100644 --- a/.agents/skills/updatefirstmate/SKILL.md +++ b/.agents/skills/updatefirstmate/SKILL.md @@ -18,6 +18,9 @@ It never forces, never creates a merge commit, never stashes, and advances a tar A tracked-files fast-forward leaves the gitignored operational dirs (data/, state/, config/, projects/, .no-mistakes/) untouched, so a secondmate's in-flight work is never disrupted. This touches only the firstmate repo and its own worktrees, never anything under `projects/`. +Post-swap, this pull follows `origin`, which is the captain's fork, so `/updatefirstmate` converges the whole fleet onto that fork and does nothing on the primary checkout, which is itself the source every fork tracks. +It deliberately never reaches for the read-only `upstream` template: folding template improvements into local `main` is a separate, reviewed reconciliation ship task, the one the `UPSTREAM_DRIFT:` bootstrap diagnostic surfaces, and it never rides in on a self-update. + ## What it does 1. **Run the updater:** diff --git a/AGENTS.md b/AGENTS.md index fdf758c1c..446113214 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -363,6 +363,8 @@ Firstmate's shared instruction surface reaches running homes only after it lands Only `AGENTS.md`, `bin/`, and `.agents/skills/` are loaded by a running firstmate; public `skills/` is an installer-facing surface. When the captain invokes `/updatefirstmate` or asks to update firstmate, load the `/updatefirstmate` skill. It performs guarded fast-forward updates of firstmate and registered secondmate homes, refreshes instructions, and never touches anything under `projects/`. +Post-swap, `origin` is the captain's fork, so `/updatefirstmate` only ever converges the fleet onto that fork and is a no-op on the primary checkout, which is itself the fleet's source of truth. +It never pulls the read-only `upstream` template; folding upstream-template improvements into local `main` is the separate reviewed reconciliation task the `UPSTREAM_DRIFT:` bootstrap line surfaces (section 3), never a self-update. ## 13. Agent-only reference skills From c4d85583a16a319136f94bc3c8ae802f03787da8 Mon Sep 17 00:00:00 2001 From: fmtest Date: Tue, 14 Jul 2026 20:25:55 -0400 Subject: [PATCH 37/41] checkpoint: recover uncommitted work after captain/crew crash Crew session for this task crashed (tmux/captain core dump). This commit preserves in-flight work from the orphaned worktree so it is not lost. A tar backup also exists under /home/orangepi/fm-recovery-backups/. --- .agents/skills/verify-trivial/SKILL.md | 46 +++++++++++++++++ AGENTS.md | 5 ++ bin/fm-brief.sh | 34 +++++++++++-- docs/scripts.md | 2 +- tests/fm-brief.test.sh | 70 ++++++++++++++++++++++++++ 5 files changed, 153 insertions(+), 4 deletions(-) create mode 100644 .agents/skills/verify-trivial/SKILL.md diff --git a/.agents/skills/verify-trivial/SKILL.md b/.agents/skills/verify-trivial/SKILL.md new file mode 100644 index 000000000..a097bcecf --- /dev/null +++ b/.agents/skills/verify-trivial/SKILL.md @@ -0,0 +1,46 @@ +--- +name: verify-trivial +description: >- + Agent-only Light verification tier for a task dispatched at the trivial (haiku/low) model/effort tier. + Use when authoring or filling in a Light-tier brief's Task section (a brief scaffolded with `bin/fm-brief.sh --light-verify`), when working a task whose brief points here, or when reviewing such a task's done report before PR-ready. + Cites `verify-rigorously` for the AC bar, its 3 verification shapes, its pass/fail matrix, and its "what wasn't verified" footer; layers 3 deltas on top: a cap of 3 acceptance criteria instead of an uncapped set, a 3-attempt cap-out that escalates the task's tier rather than failing it outright, and a mandatory side-effect confidence gate before shipping. +user-invocable: false +metadata: + internal: true +--- + +# verify-trivial + +Light verification tier for a task dispatched at the trivial (haiku/low) model/effort tier (see `AGENTS.md` section 7's escalation-trigger paragraph). +Such a task is scaffolded with `bin/fm-brief.sh --light-verify`; its brief points here instead of the default, uncapped verification workflow. + +## Cite, don't copy + +Load `verify-rigorously` and apply it in full: named observable pass/fail acceptance criteria, one of its 3 verification shapes per criterion, the criterion x shape pass/fail matrix, and the closing "what wasn't verified" footer. +This skill does not restate that contract - `verify-rigorously` is its sole owner. +Everything below is a delta on top of it that applies only to a Light-tier task. + +## Delta 1: cap acceptance criteria at 3 + +Name at most 3 acceptance criteria, not verify-rigorously's uncapped set. +A trivial-tier task is scoped small enough by definition (`bin/fm-tier-guard.sh`'s envelope: at most 2 changed files, at most 30 changed lines) that 3 observable criteria should describe "done" completely. +If the task genuinely needs more than 3 to state what done means, that is itself a signal the task has outgrown the trivial tier: report it per Delta 2 instead of naming a 4th criterion. + +## Delta 2: a 3-attempt cap-out escalates, it does not fail + +verify-rigorously's own 3-attempt cap says: stop and report what's failing, what was tried, and the likely cause. +On the Light tier, that report is not a terminal failure - it is a request to escalate the task off this tier. +Append `needs-decision: {criterion} failed verification after 3 attempts, requesting escalation to a heavier tier` (or `blocked:` if the same obstacle recurred, per the brief's own rule for that case) and stop; do not report `failed:` for a cap-out alone. +Firstmate escalates the task's model/effort in place per `AGENTS.md` section 7's escalation-trigger paragraph, exactly as a `bin/fm-tier-guard.sh` envelope escalation would, and the task continues under the ordinary uncapped `verify-rigorously` workflow from there. +Never drop the task back to the Light tier afterward - this mirrors section 7's "never silently de-escalate for the rest of that task's life." + +## Delta 3: mandatory side-effect confidence gate before shipping + +Before reporting `done:` (ship tasks) or writing the report (scout tasks), state explicitly, in one line, whether the change has any side effects outside the files it touches - a shared config default, a migration, a public interface, a cross-task or cross-project coupling - and your confidence in that assessment. +An unstated "no side effects" is not the same as a stated one: silence here is exactly the gap verify-rigorously's own footer warns against, so this gate is mandatory even when the answer is "none found." +If confidence is anything less than high, that uncertainty is itself grounds for a Delta 2 escalation rather than shipping on the Light tier. + +## Wiring this tier's no-mistakes validation + +For a no-mistakes-mode ship task on this tier, firstmate's validation trigger skips the document step (`no-mistakes axi run --skip=document`): the side-effect confidence gate above already covers what that step would otherwise catch at this tier's scale. +This is firstmate's own action at Validate time (`AGENTS.md` section 7), not something the crewmate invokes itself. diff --git a/AGENTS.md b/AGENTS.md index 446113214..f4b3d8bae 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -221,6 +221,9 @@ A report escalates the task to at least Sonnet/high in place, without losing its A crewmate's own report - it cannot find root cause, a "confirmed" fix touches a shared path, or it raised a tradeoff buried in seemingly mechanical work - is a second, non-mechanical trigger with the same effect. Either way, escalate the model/effort in place and never silently de-escalate for the rest of that task's life. +A trivial-tier task also runs the Light verification tier: scaffold its brief with `bin/fm-brief.sh --light-verify` (section 11) so it cites `verify-trivial` instead of the default full verification workflow, and for a no-mistakes-mode task, skip the document step when triggering its validation (`no-mistakes axi run --skip=document`) since that tier's mandatory side-effect confidence gate already covers what the step would check. +A tier-guard or self-reported escalation above lifts the task off the Light verification tier for the rest of its life, exactly like the model/effort escalation it rides alongside. + **Ultracode confirmation.** Before advancing an ultracode-flagged task to PR-ready, run `bin/fm-ultracode-guard.sh check `; it refuses until a genuinely separate task - dispatched independently, never a sub-task the flagged crewmate spawned itself - has reviewed the finished diff and its findings were addressed, recorded with `bin/fm-ultracode-guard.sh reviewed `. @@ -352,6 +355,7 @@ Every ship brief must retain the worktree-isolation assertion and stop if launch If a ship task touches firstmate's shared tracked material, explicitly require `firstmate-coding-guidelines` before editing. If a task will drive Herdr lifecycle behavior, scaffold with `--herdr-lab`; if that need appears after an unguarded scaffold, stop and regenerate rather than adding commands by hand. The generated Herdr contract must use a named non-`default` isolated lab and its guarded helper for every lifecycle action. +For a task dispatched at the trivial (Haiku/low) tier, scaffold with `--light-verify` so the brief points at `verify-trivial` (loaded per section 13's trigger) instead of the default full verification workflow; the flag is likewise rejected for `--secondmate` briefs. Load `secondmate-provisioning` before creating or using a charter brief and preserve its idle-by-default and marked-return-channel contracts. Status appends are sparse supervisor-actionable events, not routine progress; `bin/fm-classify-lib.sh` owns keyed open and resolved semantics. @@ -380,6 +384,7 @@ These skills are not captain-invocable; load them only at their precise triggers - `fmx-respond` - load on an `x-mention ` `check:` wake to handle the mention, on an `x-mode-error ...` `check:` wake to report the X-mode configuration blocker, and on any milestone or terminal wake for an X-mode-linked task before posting its completion follow-up; relevant only when X mode is on. - `firstmate-codexapp` - load before coordinating a visible Codex Desktop thread, evaluating a Codex App backend request, or reconciling Codex Desktop host-tool smoke evidence for Firstmate work. - `firstmate-coding-guidelines` - load before changing firstmate's shared, tracked material, as defined by section 1's list, whether editing directly or briefing a crewmate for a firstmate-repo task. +- `verify-trivial` - load before authoring or filling in a Light-tier brief's Task section (a brief scaffolded with `bin/fm-brief.sh --light-verify`, section 11) and before advancing such a task to PR-ready. ## 14. X mode diff --git a/bin/fm-brief.sh b/bin/fm-brief.sh index c3ffab689..c16c81fe9 100755 --- a/bin/fm-brief.sh +++ b/bin/fm-brief.sh @@ -6,7 +6,7 @@ # description, acceptance criteria, and context, and may adjust other sections # when the task genuinely deviates (e.g. working an existing external PR instead # of shipping a new one). -# Usage: fm-brief.sh [--scout] [--herdr-lab] +# Usage: fm-brief.sh [--scout] [--herdr-lab] [--light-verify] # fm-brief.sh --secondmate {...|--no-projects} # --scout writes the scout contract instead: the deliverable is a report at # data//report.md (no branch, no push, no PR) and the worktree is scratch. @@ -26,6 +26,13 @@ # The flag must be explicit because {TASK} is filled after scaffolding and the # caller-supplied repo string cannot reliably identify this repo. Briefs made # without it carry a loud declaration so an omitted contract cannot be silent. +# --light-verify marks a ship or scout brief for the Light verification tier +# (see AGENTS.md section 7 and the verify-trivial skill): firstmate decides to +# pass this when the task is dispatched at the trivial (haiku/low) tier. It +# points the crewmate at verify-trivial instead of the default full +# verification workflow and, for a no-mistakes-mode ship task, notes that +# firstmate's validation trigger will skip the document step. Rejected for +# --secondmate briefs, which are never tier-dispatched. # For ship tasks, the definition of done is shaped by the project's delivery mode # (data/projects.md via fm-project-mode.sh; see the project-management skill # and AGENTS.md task lifecycle): @@ -76,6 +83,7 @@ STATE="${FM_STATE_OVERRIDE:-$FM_HOME/state}" KIND=ship HERDR_LAB=0 NO_PROJECTS=0 +LIGHT_VERIFY=0 POS=() for a in "$@"; do case "$a" in @@ -83,6 +91,7 @@ for a in "$@"; do --secondmate) KIND=secondmate ;; --herdr-lab) HERDR_LAB=1 ;; --no-projects) NO_PROJECTS=1 ;; + --light-verify) LIGHT_VERIFY=1 ;; *) POS+=("$a") ;; esac done @@ -93,6 +102,11 @@ if [ "$KIND" = secondmate ] && [ "$HERDR_LAB" -eq 1 ]; then exit 1 fi +if [ "$KIND" = secondmate ] && [ "$LIGHT_VERIFY" -eq 1 ]; then + echo "error: --light-verify applies only to crewmate ship or scout briefs" >&2 + exit 1 +fi + if [ "$NO_PROJECTS" -eq 1 ] && [ "$KIND" != secondmate ]; then echo "error: --no-projects applies only to --secondmate charters" >&2 exit 1 @@ -223,6 +237,19 @@ EOF ) fi +if [ "$LIGHT_VERIFY" -eq 1 ]; then +VERIFY_TRIVIAL_PATH=$(shell_quote "$FM_ROOT/.agents/skills/verify-trivial/SKILL.md") +VERIFY_TIER_SECTION=$(printf '%s\n' \ +'# Verification tier: Light' \ +'This brief was scaffolded with `--light-verify`: read '"$VERIFY_TRIVIAL_PATH"' for the acceptance-criteria cap, the escalate-not-fail cap-out, and the mandatory side-effect confidence gate that apply to this task instead of the default full verification workflow.' \ +'That file itself points to `verify-rigorously` (load it normally) for the AC bar, verification shapes, pass/fail matrix, and "what wasn'\''t verified" footer this tier still applies in full.') +VERIFY_TIER_BLOCK=$'\n'"$VERIFY_TIER_SECTION"$'\n' +LIGHT_VERIFY_DOD_NOTE=$'\n'"This task runs the Light verification tier (see the Verification tier section above): when firstmate sends the validation trigger for this task, it skips the document step, since the side-effect confidence gate already covers what that step would check at this scale." +else +VERIFY_TIER_BLOCK="" +LIGHT_VERIFY_DOD_NOTE="" +fi + if [ "$KIND" = scout ]; then cat > "$BRIEF" </dev/null 2>&1 + brief="$home/data/$id/brief.md" + assert_present "$brief" "Light-verify scout brief was not scaffolded" + assert_grep "# Verification tier: Light" "$brief" \ + "Light-verify brief missing its Verification tier heading" + assert_grep "$ROOT/.agents/skills/verify-trivial/SKILL.md" "$brief" \ + "Light-verify brief missing the absolute verify-trivial pointer" + assert_grep "verify-rigorously" "$brief" \ + "Light-verify brief lost its verify-rigorously cross-reference" + pass "fm-brief.sh: --light-verify points the brief at verify-trivial" +} + +test_light_verify_absent_by_default() { + local home id brief + home="$TMP_ROOT/no-light-verify-home" + mkdir -p "$home/data" + id="brief-no-light-verify-e2" + FM_HOME="$home" "$ROOT/bin/fm-brief.sh" "$id" firstmate >/dev/null 2>&1 + brief="$home/data/$id/brief.md" + assert_no_grep "# Verification tier: Light" "$brief" \ + "brief scaffolded without --light-verify should not mention the Light verification tier" + assert_no_grep "skips the document step" "$brief" \ + "brief scaffolded without --light-verify should not carry the skip-document note" + pass "fm-brief.sh: omitting --light-verify leaves no Light-tier trace" +} + +test_light_verify_skips_document_for_nomistakes_mode_only() { + local home brief + home="$TMP_ROOT/light-verify-dod-home" + mkdir -p "$home/data" + cat > "$home/data/projects.md" <<'EOF' +- direct-proj [direct-PR] - fixture for direct-PR mode (added 2026-07-01) +EOF + + FM_HOME="$home" "$ROOT/bin/fm-brief.sh" brief-light-nomistakes-e3 no-registry-proj --light-verify >/dev/null 2>&1 + brief="$home/data/brief-light-nomistakes-e3/brief.md" + assert_grep "skips the document step" "$brief" \ + "no-mistakes-mode Light-verify brief missing the skip-document note" + + FM_HOME="$home" "$ROOT/bin/fm-brief.sh" brief-light-directpr-e4 direct-proj --light-verify >/dev/null 2>&1 + brief="$home/data/brief-light-directpr-e4/brief.md" + assert_grep "# Verification tier: Light" "$brief" \ + "direct-PR Light-verify brief should still get the Task-section pointer" + assert_no_grep "skips the document step" "$brief" \ + "direct-PR mode never runs no-mistakes, so it should not carry the skip-document note" + + pass "fm-brief.sh: --light-verify's skip-document note is scoped to no-mistakes-mode ship tasks" +} + +test_light_verify_rejected_for_secondmate() { + local home status=0 + home="$TMP_ROOT/light-verify-secondmate-home" + mkdir -p "$home/data" + FM_HOME="$home" FM_SECONDMATE_CHARTER=x \ + "$ROOT/bin/fm-brief.sh" sm-light-verify-e5 --secondmate --no-projects --light-verify >/dev/null 2>&1 || status=$? + expect_code 1 "$status" "secondmate --light-verify must be rejected" + assert_absent "$home/data/sm-light-verify-e5/brief.md" \ + "rejected secondmate --light-verify still wrote a brief" + pass "fm-brief.sh: --light-verify is rejected for secondmate charters" +} + test_pause_verb_override_renders_all_brief_scaffolds() { local home kind id brief home="$TMP_ROOT/pause-verb-home" @@ -300,4 +366,8 @@ test_herdr_lab_contract_quotes_foreign_firstmate_path test_herdr_lab_omission_is_loud_for_ship_and_scout test_herdr_lab_contract_applies_to_scouts_but_not_secondmates test_secondmate_no_projects_charter +test_light_verify_points_to_verify_trivial +test_light_verify_absent_by_default +test_light_verify_skips_document_for_nomistakes_mode_only +test_light_verify_rejected_for_secondmate test_pause_verb_override_renders_all_brief_scaffolds From 65a4deab557d33b22ce453b55ebc2fee243d3dab Mon Sep 17 00:00:00 2001 From: fmtest Date: Tue, 14 Jul 2026 23:18:47 -0400 Subject: [PATCH 38/41] fix: suppress SC2016 for the deliberate single-quoted verify-tier brief text --- bin/fm-brief.sh | 1 + 1 file changed, 1 insertion(+) diff --git a/bin/fm-brief.sh b/bin/fm-brief.sh index c16c81fe9..c214da5f3 100755 --- a/bin/fm-brief.sh +++ b/bin/fm-brief.sh @@ -239,6 +239,7 @@ fi if [ "$LIGHT_VERIFY" -eq 1 ]; then VERIFY_TRIVIAL_PATH=$(shell_quote "$FM_ROOT/.agents/skills/verify-trivial/SKILL.md") +# shellcheck disable=SC2016 # single quotes are deliberate: this is literal brief text whose backtick-wrapped names must reach the reading agent verbatim, not expand at scaffold time; only the '"$VAR"' break-outs interpolate. VERIFY_TIER_SECTION=$(printf '%s\n' \ '# Verification tier: Light' \ 'This brief was scaffolded with `--light-verify`: read '"$VERIFY_TRIVIAL_PATH"' for the acceptance-criteria cap, the escalate-not-fail cap-out, and the mandatory side-effect confidence gate that apply to this task instead of the default full verification workflow.' \ From c334e20e5a78303c235c499b8238f9c4cf62ab89 Mon Sep 17 00:00:00 2001 From: fmtest Date: Wed, 15 Jul 2026 01:58:40 -0400 Subject: [PATCH 39/41] no-mistakes(review): harden guardrail and deploy-sync fail-open paths per review --- bin/fm-bootstrap.sh | 8 +++++ bin/fm-nas-deploy-sync.sh | 56 ++++++++++++++++++++++++-------- bin/fm-risk-tripwire.sh | 49 ++++++++++++++++++---------- bin/fm-ultracode-guard.sh | 13 ++++++++ tests/fm-backend-herdr.test.sh | 9 ----- tests/fm-bootstrap.test.sh | 4 +++ tests/fm-brief.test.sh | 1 + tests/fm-nas-deploy-sync.test.sh | 22 +++++++++++++ tests/fm-teardown.test.sh | 20 ------------ 9 files changed, 122 insertions(+), 60 deletions(-) diff --git a/bin/fm-bootstrap.sh b/bin/fm-bootstrap.sh index 0ee362b5d..6e2a85e21 100755 --- a/bin/fm-bootstrap.sh +++ b/bin/fm-bootstrap.sh @@ -626,6 +626,14 @@ crew_dispatch_validate() { "unknown select: " + ([ (.rules // [])[]? | .select? // empty | select(. != "quota-balanced") ] | unique | join(", ")) elif has("default") and (.default | type) != "object" then "default must be an object" elif has("default") and ((.default.harness? | type) != "string" or (.default.harness | length) == 0) then "default needs harness when present" + elif ([(.rules // [])[]? | use_profiles(.use?)[]?] + + (if (.default? | type) == "object" then [.default] else [] end)) + | map(select(has("ultracode") and (.ultracode | type) != "boolean")) | length > 0 + then "ultracode must be a JSON boolean (fm-dispatch-select.sh silently drops any other type)" + elif ([(.rules // [])[]? | use_profiles(.use?)[]?] + + (if (.default? | type) == "object" then [.default] else [] end)) + | map(select(has("ultracode_role") and ((.ultracode_role | type) != "string" or (.ultracode_role | length) == 0))) | length > 0 + then "ultracode_role must be a non-empty string (fm-dispatch-select.sh silently drops any other type)" else ([(.rules // [])[]? | use_profiles(.use?)[]?.harness] + [.default?.harness?] | map(select(. != null)) diff --git a/bin/fm-nas-deploy-sync.sh b/bin/fm-nas-deploy-sync.sh index 1233dc9f7..b4a25a50e 100755 --- a/bin/fm-nas-deploy-sync.sh +++ b/bin/fm-nas-deploy-sync.sh @@ -28,11 +28,12 @@ # ahead of it on PATH intercepts restart/list during tests - real tests must never # reach the real /mnt/nas/experiments or a real pm2 process. # -# Every filesystem/git touch of $NAS_PATH is wrapped in a bounded timeout -# (FM_NAS_SYNC_TIMEOUT seconds, default 15): the mount is a NAS, and an -# unreachable one is a hang risk (a stuck stat/read), not a fast failure, which -# would otherwise stall fm-teardown.sh's caller and break this script's -# non-blocking contract. +# Every filesystem/git touch of $NAS_PATH - and the pm2 restart/jlist calls, +# whose target process's script lives on the same mount - is wrapped in a +# bounded timeout (FM_NAS_SYNC_TIMEOUT seconds, default 15, SIGKILL 5s after +# SIGTERM): the mount is a NAS, and an unreachable one is a hang risk (a stuck +# stat/read), not a fast failure, which would otherwise stall fm-teardown.sh's +# caller and break this script's non-blocking contract. # # $NAS_PATH is one single shared live-deployment checkout - unlike a project's # projects/ dev clone, which is scoped to one firstmate home, two ship @@ -50,6 +51,10 @@ FM_HOME="${FM_HOME:-${FM_ROOT_OVERRIDE:-$FM_ROOT}}" DATA="${FM_DATA_OVERRIDE:-$FM_HOME/data}" MAP="${FM_NAS_DEPLOYMENTS_OVERRIDE:-$DATA/nas-deployments.md}" NAS_SYNC_TIMEOUT="${FM_NAS_SYNC_TIMEOUT:-15}" +if ! [[ "$NAS_SYNC_TIMEOUT" =~ ^([0-9]+([.][0-9]*)?|[.][0-9]+)$ ]]; then + echo "nas-deploy-sync: invalid FM_NAS_SYNC_TIMEOUT '$NAS_SYNC_TIMEOUT'; using 15s" >&2 + NAS_SYNC_TIMEOUT=15 +fi # shellcheck source=bin/fm-lock-lib.sh . "$SCRIPT_DIR/fm-lock-lib.sh" FM_LOCK_LOG_PREFIX=nas-deploy-sync @@ -72,11 +77,14 @@ elif command -v gtimeout >/dev/null 2>&1; then fi # bounded : run , killed after $NAS_SYNC_TIMEOUT seconds when -# a `timeout`/`gtimeout` binary is available, run directly otherwise. +# a `timeout`/`gtimeout` binary is available, run directly otherwise. -k gives +# SIGKILL 5s after SIGTERM: a process wedged in an uninterruptible NAS wait can +# ignore SIGTERM, and without the follow-up kill it would stall this script's +# caller (fm-teardown.sh) indefinitely. bounded() { case "$HAVE_TIMEOUT" in - timeout) timeout "$NAS_SYNC_TIMEOUT" "$@" ;; - gtimeout) gtimeout "$NAS_SYNC_TIMEOUT" "$@" ;; + timeout) timeout -k 5 "$NAS_SYNC_TIMEOUT" "$@" ;; + gtimeout) gtimeout -k 5 "$NAS_SYNC_TIMEOUT" "$@" ;; *) "$@" ;; esac } @@ -183,7 +191,10 @@ lookup_row() { ' "$MAP" } -row=$(lookup_row) +row=$(lookup_row) || { + echo "$NAME: skipped: cannot read deployment mapping $MAP" + exit 0 +} if [ -z "$row" ]; then echo "$NAME: skipped: no recorded NAS deployment" exit 0 @@ -238,8 +249,14 @@ if ! git_nas rev-parse --verify --quiet "$BASE^{commit}" >/dev/null; then fi cur=$(git_nas symbolic-ref --quiet --short HEAD 2>/dev/null || echo "") +# A failed/timed-out status read must not read as "clean" - that is exactly the +# unreachable-NAS case where merging a possibly dirty checkout would be wrong. +status_out=$(git_nas status --porcelain 2>/dev/null) || { + echo "$NAME: skipped: cannot read git status at $NAS_PATH" + exit 0 +} dirty=no -[ -z "$(git_nas status --porcelain 2>/dev/null | head -1)" ] || dirty=yes +[ -z "$status_out" ] || dirty=yes if [ "$cur" != "$DEFAULT" ]; then echo "$NAME: STUCK: NAS checkout at $NAS_PATH is not on $DEFAULT - needs attention" @@ -270,7 +287,10 @@ if ! git_nas merge-base --is-ancestor "$DEFAULT" "$BASE"; then exit 0 fi -before=$(git_nas rev-parse --short "$DEFAULT") +before=$(git_nas rev-parse --short "$DEFAULT") || { + echo "$NAME: skipped: cannot read $DEFAULT at $NAS_PATH" + exit 0 +} merge_output="" if ! merge_output=$(git_nas merge --ff-only "$BASE" 2>&1); then reason="fast-forward failed" @@ -278,14 +298,16 @@ if ! merge_output=$(git_nas merge --ff-only "$BASE" 2>&1); then echo "$NAME: skipped: $reason" exit 0 fi -after=$(git_nas rev-parse --short "$DEFAULT") +# The merge already advanced the checkout; a transient failure reading the new +# rev must not abort before the pm2 restart below runs. +after=$(git_nas rev-parse --short "$DEFAULT" 2>/dev/null) || after="?" # process_online : true when pm2 currently reports EVERY instance of # as online (a clustered app can register several jlist entries under # the same name; one lagging instance must not be masked by the others). process_online() { local proc=$1 status - status=$(pm2 jlist 2>/dev/null | jq -r --arg n "$proc" \ + status=$(bounded pm2 jlist 2>/dev/null | jq -r --arg n "$proc" \ '[.[] | select(.name == $n) | .pm2_env.status] as $s | if ($s | length) == 0 then "missing" elif all($s[]; . == "online") then "online" @@ -295,12 +317,14 @@ process_online() { } restart_problems="" +restarted=0 IFS=',' read -r -a procs <<< "$PROCS" for proc in "${procs[@]}"; do proc="${proc#"${proc%%[![:space:]]*}"}" proc="${proc%"${proc##*[![:space:]]}"}" [ -n "$proc" ] || continue - if ! pm2 restart "$proc" >/dev/null 2>&1; then + restarted=$(( restarted + 1 )) + if ! bounded pm2 restart "$proc" >/dev/null 2>&1; then restart_problems="$restart_problems $proc(restart failed)" continue fi @@ -309,6 +333,10 @@ for proc in "${procs[@]}"; do fi done +if [ "$restarted" -eq 0 ]; then + echo "$NAME: synced $before..$after but no pm2 processes recorded in the mapping row - nothing restarted, needs attention" + exit 0 +fi if [ -n "$restart_problems" ]; then echo "$NAME: synced $before..$after but restart verification failed:$restart_problems - needs attention" exit 0 diff --git a/bin/fm-risk-tripwire.sh b/bin/fm-risk-tripwire.sh index e684eff57..84b724b94 100755 --- a/bin/fm-risk-tripwire.sh +++ b/bin/fm-risk-tripwire.sh @@ -18,8 +18,9 @@ # 0 no risk signal found # 1 a RISK hit fired (one "RISK: " line per hit is printed) # 2 could not check: a malformed invocation, neither a brief nor a usable -# worktree/project exists yet (nothing to check), or the worktree/project -# exist but their diff base could not be resolved, so the binding diff +# worktree/project exists yet (nothing to check), or the meta records a +# worktree/project whose directories vanished, whose diff base could not +# be resolved, or whose diff command itself failed, so the binding diff # checkpoint could not run (warned to stderr, never a silent clean pass) # Distinct codes matter so a caller branching on $? cannot mistake a malformed # invocation for a real risk hit. A hit means: floor this task's model/effort to @@ -186,7 +187,13 @@ META="$STATE/$ID.meta" if [ -f "$META" ]; then WT=$(grep '^worktree=' "$META" | tail -1 | cut -d= -f2- || true) PROJ=$(grep '^project=' "$META" | tail -1 | cut -d= -f2- || true) - if [ -n "$WT" ] && [ -n "$PROJ" ] && [ -d "$WT" ] && [ -d "$PROJ" ]; then + if [ -n "$WT" ] && [ -n "$PROJ" ] && { [ ! -d "$WT" ] || [ ! -d "$PROJ" ]; }; then + # The meta promises a diff checkpoint (worktree=/project= recorded), but the + # directories are gone - a vanished worktree at Validate time must not read + # as a clean pass, exactly like an unresolvable diff base below. + echo "warning: meta for $ID records worktree=$WT and project=$PROJ but at least one no longer exists as a directory; the diff checkpoint did not run - reporting not-checkable (exit 2), not a clean pass" >&2 + DIFF_UNRESOLVED=1 + elif [ -n "$WT" ] && [ -n "$PROJ" ]; then CHECKED=1 diff_base_resolved=0 # The base resolution below duplicates bin/fm-review-diff.sh's origin fetch @@ -204,20 +211,28 @@ if [ -f "$META" ]; then fi if git -C "$WT" rev-parse --verify --quiet "$BASE^{commit}" >/dev/null 2>&1; then diff_base_resolved=1 - diff_paths=$(git -C "$WT" diff --name-only "$BASE...HEAD" -- 2>/dev/null || true) - risky_paths= - if [ -n "$diff_paths" ]; then - while IFS= read -r path; do - [ -n "$path" ] || continue - lower_path=$(printf '%s' "$path" | tr '[:upper:]' '[:lower:]') - if path_is_risky "$lower_path"; then - risky_paths="${risky_paths}${risky_paths:+, }$path" - fi - done <<< "$diff_paths" - fi - if [ -n "$risky_paths" ]; then - echo "RISK: diff for $ID touches risk-adjacent path(s): $risky_paths" - FOUND=1 + # A failing diff (no merge base after a rebase onto unrelated history, + # a shallow clone) must not read as an empty path list: that would be + # the silent clean pass the not-checkable contract below exists to + # prevent. + if diff_paths=$(git -C "$WT" diff --name-only "$BASE...HEAD" -- 2>/dev/null); then + risky_paths= + if [ -n "$diff_paths" ]; then + while IFS= read -r path; do + [ -n "$path" ] || continue + lower_path=$(printf '%s' "$path" | tr '[:upper:]' '[:lower:]') + if path_is_risky "$lower_path"; then + risky_paths="${risky_paths}${risky_paths:+, }$path" + fi + done <<< "$diff_paths" + fi + if [ -n "$risky_paths" ]; then + echo "RISK: diff for $ID touches risk-adjacent path(s): $risky_paths" + FOUND=1 + fi + else + echo "warning: git diff --name-only $BASE...HEAD failed for $ID in $WT; the diff checkpoint did not run - reporting not-checkable (exit 2), not a clean pass" >&2 + DIFF_UNRESOLVED=1 fi fi fi diff --git a/bin/fm-ultracode-guard.sh b/bin/fm-ultracode-guard.sh index 485510fa3..24c94964b 100755 --- a/bin/fm-ultracode-guard.sh +++ b/bin/fm-ultracode-guard.sh @@ -50,6 +50,18 @@ if [ -z "$CMD" ] || [ -z "$ID" ]; then exit 1 fi +# Same unsafe-id shape check as bin/fm-x-link.sh: task ids are plain slugs, so +# a path-shaped id ("./fix-x", "x/../fix-x") must be rejected before it is +# compared against another id or used to build a state/ path - otherwise a +# task can smuggle itself in as its own "independent" reviewer, or point the +# meta probe outside state/. +validate_task_id() { + case "$2" in + ''|.*|*[!A-Za-z0-9._-]*) echo "error: unsafe $1: $2" >&2; exit 1 ;; + esac +} +validate_task_id task-id "$ID" + MARKER="$STATE/$ID.ultracode" cmd_flag() { @@ -67,6 +79,7 @@ cmd_reviewed() { local reviewer=${3:-} [ $# -eq 3 ] || { usage; exit 1; } [ -n "$reviewer" ] || { echo "error: reviewed requires a reviewer-task-id" >&2; exit 1; } + validate_task_id reviewer-task-id "$reviewer" [ -f "$MARKER" ] || { echo "error: $ID is not ultracode-flagged (no $MARKER); nothing to mark reviewed" >&2; exit 1; } if [ "$reviewer" = "$ID" ]; then echo "error: reviewer-task-id must be a task distinct from $ID - a task cannot independently review itself" >&2 diff --git a/tests/fm-backend-herdr.test.sh b/tests/fm-backend-herdr.test.sh index 77484a693..e099b6b97 100755 --- a/tests/fm-backend-herdr.test.sh +++ b/tests/fm-backend-herdr.test.sh @@ -51,15 +51,6 @@ if [ "${1:-}" = status ] && [ "${2:-}" = --json ] && [ "${FM_HERDR_SCRIPT_STATUS printf '{"client":{"version":"0.7.1","protocol":14},"server":{"running":true}}\n' exit 0 fi -# fm_backend_herdr_server_ensure launches `herdr server` fire-and-forget -# (backgrounded) so its poll loop can proceed concurrently. No test scripts a -# canned response for it, so it must never read/write the shared call-ordinal -# COUNT_FILE below - doing so races with whichever synchronous call the poll -# loop issues at the same moment, losing an update and permanently shifting -# every later call's response number. -if [ "${1:-}" = server ]; then - exit 0 -fi next=$(( $(cat "$COUNT_FILE" 2>/dev/null || echo 0) + 1 )) n=$next echo "$n" > "$COUNT_FILE" diff --git a/tests/fm-bootstrap.test.sh b/tests/fm-bootstrap.test.sh index 779c31d5c..5754ea037 100755 --- a/tests/fm-bootstrap.test.sh +++ b/tests/fm-bootstrap.test.sh @@ -728,6 +728,10 @@ empty array use is flagged^{"rules":[{"when":"big feature","use":[]}]}^exact^CRE array profile without harness is flagged^{"rules":[{"when":"big feature","use":[{"model":"gpt-5.5"}]}]}^exact^CREW_DISPATCH: invalid config/crew-dispatch.json - each use profile needs harness unknown select is flagged^{"rules":[{"when":"big feature","use":[{"harness":"claude"},{"harness":"codex"}],"select":"mystery"}]}^exact^CREW_DISPATCH: invalid config/crew-dispatch.json - unknown select: mystery array profile unsupported effort is flagged^{"rules":[{"when":"big feature","use":[{"harness":"codex","effort":"max"}]}]}^exact^CREW_DISPATCH: invalid config/crew-dispatch.json - invalid effort: codex:max +mistyped ultracode string is flagged^{"rules":[{"when":"risky work","use":{"harness":"claude","ultracode":"true"}}]}^exact^CREW_DISPATCH: invalid config/crew-dispatch.json - ultracode must be a JSON boolean (fm-dispatch-select.sh silently drops any other type) +mistyped default ultracode number is flagged^{"rules":[{"when":"x","use":{"harness":"claude"}}],"default":{"harness":"claude","ultracode":1}}^exact^CREW_DISPATCH: invalid config/crew-dispatch.json - ultracode must be a JSON boolean (fm-dispatch-select.sh silently drops any other type) +mistyped ultracode_role is flagged^{"rules":[{"when":"risky work","use":{"harness":"claude","ultracode":true,"ultracode_role":1}}]}^exact^CREW_DISPATCH: invalid config/crew-dispatch.json - ultracode_role must be a non-empty string (fm-dispatch-select.sh silently drops any other type) +valid ultracode profile is accepted^{"rules":[{"when":"risky work","use":{"harness":"claude","ultracode":true,"ultracode_role":"independent-review"}}]}^grep^CREW_DISPATCH: active config/crew-dispatch.json ROWS pass "bootstrap validates crew-dispatch.json and reports malformed or unverified configs" } diff --git a/tests/fm-brief.test.sh b/tests/fm-brief.test.sh index 62694a019..5f17d4a76 100755 --- a/tests/fm-brief.test.sh +++ b/tests/fm-brief.test.sh @@ -277,6 +277,7 @@ test_light_verify_absent_by_default() { id="brief-no-light-verify-e2" FM_HOME="$home" "$ROOT/bin/fm-brief.sh" "$id" firstmate >/dev/null 2>&1 brief="$home/data/$id/brief.md" + assert_present "$brief" "default brief was not scaffolded" assert_no_grep "# Verification tier: Light" "$brief" \ "brief scaffolded without --light-verify should not mention the Light verification tier" assert_no_grep "skips the document step" "$brief" \ diff --git a/tests/fm-nas-deploy-sync.test.sh b/tests/fm-nas-deploy-sync.test.sh index ecffe6e29..c9f66ce57 100755 --- a/tests/fm-nas-deploy-sync.test.sh +++ b/tests/fm-nas-deploy-sync.test.sh @@ -303,6 +303,27 @@ test_clustered_partial_restart_is_not_masked_online() { pass "a clustered pm2 app with one lagging instance is reported as a restart problem, not masked as online" } +test_empty_procs_row_is_not_reported_verified() { + local home nas log out + home=$(new_home) + nas=$(build_nas_pair "$home" noprocs-test) + advance_origin "$home" noprocs-test C1 + write_map "$home" noprocs-test "$nas" "" + log="$home/restart.log" + write_pm2_stub "$home/fakebin" "$log" + + out=$(run_sync "$home" noprocs-test) + + assert_contains "$out" "no pm2 processes recorded" \ + "empty-procs: did not report the missing pm2 mapping" + assert_contains "$out" "needs attention" "empty-procs: did not flag the deployment as needing attention" + case "$out" in + *"verified online"*) fail "empty-procs: a row with no pm2 processes was reported as verified online" ;; + esac + [ ! -s "$log" ] || fail "empty-procs: pm2 restart was invoked despite an empty process column" + pass "a mapping row with an empty pm2 column syncs but reports no processes to restart, never verified online" +} + test_hung_fetch_is_bounded_by_timeout() { local home nas out rc start elapsed if ! command -v timeout >/dev/null 2>&1 && ! command -v gtimeout >/dev/null 2>&1; then @@ -466,6 +487,7 @@ test_clean_behind_fast_forwards_and_restarts test_dirty_is_skipped_untouched test_diverged_is_stuck_untouched test_clustered_partial_restart_is_not_masked_online +test_empty_procs_row_is_not_reported_verified test_absent_project_is_silent_noop test_hung_fetch_is_bounded_by_timeout test_transient_packed_refs_lock_is_retried diff --git a/tests/fm-teardown.test.sh b/tests/fm-teardown.test.sh index bfe0f8ee4..b49a5d778 100755 --- a/tests/fm-teardown.test.sh +++ b/tests/fm-teardown.test.sh @@ -209,18 +209,6 @@ land_on_origin_main() { rm -rf "$tmp" } -# content_in_default (bin/fm-teardown.sh) relies on `git merge-tree -# --write-tree`, added in git 2.38. Probe the REAL capability rather than -# guessing a version cutoff: older git treats the flag as an unrecognized -# revision ("unknown rev --write-tree"), which the two tests exercising that -# fallback below cannot pass under. -fm_git_supports_merge_tree_write_tree() { - case "$(git -C "$ROOT" merge-tree --write-tree HEAD HEAD 2>&1)" in - *"unknown rev"*) return 1 ;; - *) return 0 ;; - esac -} - # Override GitHub lookups to report PR 7 as merged with the supplied head. add_gh_pr_merged_for_head() { local case_dir=$1 head=$2 @@ -911,10 +899,6 @@ test_pr_check_records_remote_head_when_local_lags() { test_content_in_default_fallback_allows() { local case_dir rc - fm_git_supports_merge_tree_write_tree || { - echo "skip: this git lacks 'merge-tree --write-tree' (needs git 2.38+)" - return 0 - } case_dir=$(make_case content-landed) write_meta "$case_dir" no-mistakes ship # No pr= recorded and the default gh-axi mock reports no PR, so the merged-PR path @@ -935,10 +919,6 @@ test_content_in_default_fallback_allows() { test_content_fallback_refreshes_stale_origin_ref() { local case_dir rc - fm_git_supports_merge_tree_write_tree || { - echo "skip: this git lacks 'merge-tree --write-tree' (needs git 2.38+)" - return 0 - } case_dir=$(make_case content-stale-ref) write_meta "$case_dir" no-mistakes ship wt_commit_file "$case_dir" feature.txt hello "add feature" From 32f0a51ce2967205e4d326f0bc04513a8c8b3a89 Mon Sep 17 00:00:00 2001 From: fmtest Date: Wed, 15 Jul 2026 17:37:04 -0400 Subject: [PATCH 40/41] no-mistakes(test): Fix AFK and stale-lock test failures --- bin/fm-afk-launch.sh | 55 +++++++++++++++++++++++++------- bin/fm-watch.sh | 6 +++- tests/fm-backlog-handoff.test.sh | 7 ++-- tests/fm-fleet-sync.test.sh | 23 +++++++++++++ 4 files changed, 76 insertions(+), 15 deletions(-) diff --git a/bin/fm-afk-launch.sh b/bin/fm-afk-launch.sh index a6ddcbcc6..091f13688 100755 --- a/bin/fm-afk-launch.sh +++ b/bin/fm-afk-launch.sh @@ -67,6 +67,8 @@ set +e fm_afk_launch_log() { printf 'fm-afk-launch: %s\n' "$*" >&2; } +fm_afk_launch_lock_created_by_us=0 + fm_afk_launch_lock_owned() { local pid expected actual [ -d "$FM_AFK_LAUNCH_LOCK" ] || return 1 @@ -81,16 +83,20 @@ fm_afk_launch_lock_acquire() { mkdir -p "$FM_AFK_LAUNCH_STATE" || return 1 for i in $(seq 1 200); do if mkdir "$FM_AFK_LAUNCH_LOCK" 2>/dev/null; then + fm_afk_launch_lock_created_by_us=1 if ! printf '%s' "$$" > "$FM_AFK_LAUNCH_LOCK/pid"; then rm -rf "$FM_AFK_LAUNCH_LOCK" + fm_afk_launch_lock_created_by_us=0 return 1 fi identity=$(fm_pid_identity "$$" 2>/dev/null) || { rm -rf "$FM_AFK_LAUNCH_LOCK" + fm_afk_launch_lock_created_by_us=0 return 1 } if [ -z "$identity" ] || ! printf '%s' "$identity" > "$FM_AFK_LAUNCH_LOCK/pid-identity"; then rm -rf "$FM_AFK_LAUNCH_LOCK" + fm_afk_launch_lock_created_by_us=0 return 1 fi return 0 @@ -120,6 +126,22 @@ fm_afk_launch_lock_release() { pid=$(cat "$FM_AFK_LAUNCH_LOCK/pid" 2>/dev/null || true) [ "$pid" = "$$" ] || return 0 rm -rf "$FM_AFK_LAUNCH_LOCK" + fm_afk_launch_lock_created_by_us=0 +} + +fm_afk_launch_signal_exit() { + local code=$1 + if [ -n "${child:-}" ]; then + kill -TERM "$child" 2>/dev/null || true + fi + if [ "${fm_afk_launch_lock_created_by_us:-0}" = 1 ]; then + rm -rf "$FM_AFK_LAUNCH_LOCK" + fm_afk_launch_lock_created_by_us=0 + else + fm_afk_launch_lock_release + fi + trap - EXIT INT TERM + exit "$code" } fm_afk_launch_usage() { @@ -602,20 +624,29 @@ fm_afk_launch_stop() { } fm_afk_launch_main() { - local result - fm_afk_launch_lock_acquire || return 1 + local result child + child= + trap 'fm_afk_launch_signal_exit 130' INT + trap 'fm_afk_launch_signal_exit 143' TERM + if ! fm_afk_launch_lock_acquire; then + trap - INT TERM + return 1 + fi trap fm_afk_launch_lock_release EXIT - trap 'exit 130' INT - trap 'exit 143' TERM - case "${1:-start}" in - start) fm_afk_launch_start ;; - start-native) fm_afk_launch_start_native ;; - stop) fm_afk_launch_stop ;; - reconcile) fm_afk_launch_reconcile ;; - -h|--help|help) fm_afk_launch_usage ;; - *) fm_afk_launch_usage >&2; return 2 ;; - esac + ( + case "${1:-start}" in + start) fm_afk_launch_start ;; + start-native) fm_afk_launch_start_native ;; + stop) fm_afk_launch_stop ;; + reconcile) fm_afk_launch_reconcile ;; + -h|--help|help) fm_afk_launch_usage ;; + *) fm_afk_launch_usage >&2; exit 2 ;; + esac + ) & + child=$! + wait "$child" result=$? + child= fm_afk_launch_lock_release || result=1 trap - EXIT INT TERM return "$result" diff --git a/bin/fm-watch.sh b/bin/fm-watch.sh index 13b895f00..e4e9aa791 100755 --- a/bin/fm-watch.sh +++ b/bin/fm-watch.sh @@ -849,7 +849,11 @@ EOF echo 0 > "$cf" rm -f "$ssf" "$ewf" task=$(window_to_task "$w" "$STATE") - if ! afk_present && status_is_paused "$(last_status_line "$STATE/$task.status")" && ! window_is_busy "$w" "$tail40"; then + if afk_present && status_is_paused "$(last_status_line "$STATE/$task.status")" && ! window_is_busy "$w" "$tail40"; then + fm_wake_append stale "$w" "stale: $w" || exit 1 + printf '%s' "$h" > "$sf" + wake "stale: $w" + elif ! afk_present && status_is_paused "$(last_status_line "$STATE/$task.status")" && ! window_is_busy "$w" "$tail40"; then case "$(pause_state_class "$w" "$task")" in paused) handle_paused_stale "$w" "$task" "$h" ;; *) clear_pause_tracking "$w" ;; diff --git a/tests/fm-backlog-handoff.test.sh b/tests/fm-backlog-handoff.test.sh index 9e10703e8..5868eaf12 100755 --- a/tests/fm-backlog-handoff.test.sh +++ b/tests/fm-backlog-handoff.test.sh @@ -8,10 +8,13 @@ set -u # shellcheck source=tests/secondmate-helpers.sh disable=SC1091 . "$(dirname "${BASH_SOURCE[0]}")/secondmate-helpers.sh" +# shellcheck source=bin/fm-tasks-axi-lib.sh disable=SC1091 +. "$ROOT/bin/fm-tasks-axi-lib.sh" # The move is delegated to `tasks-axi mv`, so this suite exercises the real -# binary. Skip cleanly when it is absent (matching the backend smoke suites). -command -v tasks-axi >/dev/null 2>&1 || { echo "skip: tasks-axi not found (required by the delegated handoff path)"; exit 0; } +# binary. Skip cleanly when it is absent or too old (matching the backend smoke +# suites and the production compatibility gate). +fm_tasks_axi_compatible || { echo "skip: tasks-axi not found or too old (required by the delegated handoff path)"; exit 0; } TMP_ROOT=$(fm_test_tmproot fm-backlog-handoff) diff --git a/tests/fm-fleet-sync.test.sh b/tests/fm-fleet-sync.test.sh index b1fcd0a38..af4221f7f 100755 --- a/tests/fm-fleet-sync.test.sh +++ b/tests/fm-fleet-sync.test.sh @@ -181,6 +181,26 @@ SH chmod +x "$1/git" } +# git shim: while a planted packed-refs.lock exists, fail fetch with Git's lock +# signature. Once the guard removes the stale lock, delegate to real git. +git_packed_refs_lock_until_removed() { + cat > "$1/git" <<'SH' +#!/usr/bin/env bash +real=${REAL_GIT_FOR_TEST:?} +dir=; is_fetch=0 +for a in "$@"; do [ "$a" = fetch ] && is_fetch=1; done +prev= +for a in "$@"; do [ "$prev" = -C ] && dir=$a; prev=$a; done +if [ "$is_fetch" = 1 ] && [ -n "$dir" ] && [ -e "$dir/.git/packed-refs.lock" ]; then + lock="$dir/.git/packed-refs.lock" + echo "error: could not delete reference refs/remotes/origin/feature: Unable to create '$lock': File exists." >&2 + exit 1 +fi +exec "$real" "$@" +SH + chmod +x "$1/git" +} + # run_sync_guarded [args...]: run fleet-sync # with the fakebin on PATH and stdout/stderr captured separately. Per-test knobs # (FM_FLEET_SYNC_PACKED_REFS_LOCK_*, GIT_FETCH_COUNTER) are read from the caller's @@ -479,6 +499,7 @@ test_orphaned_stale_packed_refs_lock_recovers() { clone=$(build_packed_prunable "$home" lockstale) plant_packed_refs_lock "$clone" lsof_no_holder "$fakebin" # provably no live holder + git_packed_refs_lock_until_removed "$fakebin" out="$home/out-lockstale"; err="$home/err-lockstale" set +e @@ -508,6 +529,7 @@ test_live_packed_refs_lock_is_never_removed() { clone=$(build_packed_prunable "$home" locklive) plant_packed_refs_lock "$clone" lsof_live_holder "$fakebin" # a live process holds the lock/.git open + git_packed_refs_lock_until_removed "$fakebin" before=$(head_sha "$clone") out="$home/out-locklive"; err="$home/err-locklive" @@ -533,6 +555,7 @@ test_live_git_cwd_in_clone_dir_blocks_removal() { fakebin="$home/fb-lockcwd"; rm -rf "$fakebin"; mkdir -p "$fakebin" clone=$(build_packed_prunable "$home" lockcwd) plant_packed_refs_lock "$clone" + git_packed_refs_lock_until_removed "$fakebin" # Nobody holds the lock file, but a live process holds the clone worktree as its # cwd - the narrow race where git closed packed-refs.lock but has not yet exited. lsof_holds_only_live_dir "$fakebin" From e83650385c23f9763ba155ef2974e58c39c21fec Mon Sep 17 00:00:00 2001 From: fmtest Date: Wed, 15 Jul 2026 19:52:25 -0400 Subject: [PATCH 41/41] no-mistakes(document): Document fm-merge-local push-back, ultracode validation, upstream-fetch timeout --- AGENTS.md | 2 +- bin/fm-merge-local.sh | 5 +++++ docs/configuration.md | 3 ++- 3 files changed, 8 insertions(+), 2 deletions(-) diff --git a/AGENTS.md b/AGENTS.md index f4b3d8bae..1296d7f56 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -190,7 +190,7 @@ The path's worker, automated gates, and captain approval remain authoritative: - **no-mistakes** runs the full pipeline through a PR, then waits for the configured merge authority. - **direct-PR** has the worker push and open a PR without the no-mistakes pipeline, then waits for the configured merge authority. -- **local-only** has the worker stop with a clean ready branch, then waits for the configured merge authority before firstmate uses the guarded fast-forward merge path. +- **local-only** has the worker stop with a clean ready branch, then waits for the configured merge authority before firstmate uses the guarded fast-forward merge path; when the project has an origin remote, `bin/fm-merge-local.sh` also best-effort pushes `main` there afterward so a push-backed local-only default branch never silently drifts behind local `main` (a push failure is reported but never fails the merge). Delivery mode and `yolo` are orthogonal. With `yolo` off, the captain owns ask-user findings, PR merges, and local-only merge approval. diff --git a/bin/fm-merge-local.sh b/bin/fm-merge-local.sh index 07e5ca50a..de669eb80 100755 --- a/bin/fm-merge-local.sh +++ b/bin/fm-merge-local.sh @@ -9,6 +9,11 @@ # auto-approves), and only as a clean fast-forward - it refuses a diverged branch # and tells you to have the crewmate rebase. See AGENTS.md prime directives, # project management, and task lifecycle. +# After the fast-forward, when the project has an origin remote it also +# best-effort pushes the default branch to origin, so a push-backed local-only +# default branch never silently drifts behind local main; a push failure is +# reported but never fails the merge, and a no-remote local-only project is +# skipped silently. # Usage: fm-merge-local.sh set -eu diff --git a/docs/configuration.md b/docs/configuration.md index 71158c3b1..1152f26fc 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -233,7 +233,7 @@ If a selected profile carries an effort value the chosen harness does not accept See [`docs/examples/crew-dispatch.json`](examples/crew-dispatch.json) for a starting point to copy into local `config/crew-dispatch.json`. When the file exists, bootstrap validates it with `jq`. Valid files produce a `CREW_DISPATCH: active config/crew-dispatch.json` block that lists each rule and prints `default:` when present. -Malformed JSON, an unverified harness, a malformed array profile, an unknown `select`, or an effort value unsupported by that harness is reported as `CREW_DISPATCH: invalid config/crew-dispatch.json - ...`; missing `jq` is reported through the normal `MISSING: jq` install-consent flow. +Malformed JSON, an unverified harness, a malformed array profile, an unknown `select`, an effort value unsupported by that harness, a non-boolean `ultracode`, or an empty/non-string `ultracode_role` is reported as `CREW_DISPATCH: invalid config/crew-dispatch.json - ...` (a wrong-typed `ultracode`/`ultracode_role` is otherwise silently dropped by `bin/fm-dispatch-select.sh`); missing `jq` is reported through the normal `MISSING: jq` install-consent flow. If no dispatch rule fits, firstmate uses the dispatch profile `default` when present, then falls back to `config/crew-harness`. Because the spawn backstop is gated by file presence, any fallback path after a missing match, validation error, or missing `jq` still passes a resolved harness explicitly until the file is fixed or removed. Secondmate homes inherit this file from the primary, so a secondmate's own crewmates apply the same dispatch profile behavior. @@ -412,6 +412,7 @@ FM_TIER_HEAVY_MIN_FILES=8 # fm-tier-guard.sh: changed-file count that e FM_WATCH_TRIAGE_LOG_MAX_BYTES=262144 # size cap for the watcher's absorbed-wake debug log FM_FLEET_SYNC_BOOTSTRAP_TIMEOUT= # optional seconds allowed for bootstrap's best-effort clone refresh; unset/blank defaults to max(20, 5 + 3 * origin-backed-project-count) FM_FLEET_PRUNE=1 # set to 0 to skip pruning local branches whose upstream is gone +FM_UPSTREAM_FETCH_TIMEOUT=20 # seconds allowed for the locked fleet-sync sweep's best-effort fetch of firstmate's own read-only upstream/main, refreshing the UPSTREAM_DRIFT bootstrap diagnostic FM_NAS_DEPLOYMENTS_OVERRIDE= # alternate data/nas-deployments.md path for fm-nas-deploy-sync.sh, mainly for tests FM_NAS_SYNC_TIMEOUT=15 # seconds allowed per filesystem/git touch of fm-nas-deploy-sync.sh's $NAS_PATH before it is killed, so an unreachable NAS mount cannot hang fm-teardown.sh's caller FM_NAS_SYNC_PACKED_REFS_LOCK_RETRIES=3 # fetch retries after fm-nas-deploy-sync.sh hits the orphaned .git/packed-refs.lock signature on the shared NAS checkout (two landed teardowns for the same project can race it)