Skip to content

config-reference.md

Latest commit

 

History

History
126 lines (107 loc) · 8.06 KB

config-reference.md

File metadata and controls

126 lines (107 loc) · 8.06 KB
Raw

Config reference — .olympus.json

The single source of per-repo policy. Validated by schema/olympus.schema.json. Every field is optional; an omitted value uses the default shown. With no file at all, the agents behave exactly as the defaults (back-compatible).

Precedence for any value: an exported env var > the config file > the built-in default.

project

Field Default Meaning
name — Human-readable project name.
default_branch main Branch the dev agent forks from and PRs target.

agents

Field Default Meaning
review_bot_login themis GitHub login (or a marker in the review body) used by auto-merge / revise to find the review bot's verdict.
dev_agent_name the dev agent Display name for the implementer (commit author, log lines).

labels

Field Default Meaning
assess agent:assess Manual re-triage trigger.
try agent:try Triage adds this on do; starts the dev agent.
skip agent:skip Mutes re-triage.
auto_agent auto-agent Marks dev-agent PRs for the auto-merge / revise gates.
discussing agent:discussing Marks an issue in an active triage discussion (verdict discuss); a reporter reply re-runs triage.
staging_soaked staging-soaked Added to a PR that cleared its staging soak — a human may merge it.
soak_failed soak-failed Added to a PR whose staging soak (deploy/health) didn't hold.

triage.gates

Field Default Meaning
max_loc 300 Diff-size ceiling for unattended auto-implementation.
max_files 10 File-count ceiling.
contained_areas one module, docs, or one workflow Joined into the "contained" gate text — the areas a change may touch and still qualify.
test_hint a unit / integration test What a deterministic test looks like here (shown to triage + dev agent).

triage.language

Value Behavior
auto (default) Reply in the same language the reporter used.
a code/name (en, zh, 日本語, …) Always reply in that language.

triage.auto_dispatch

When a do verdict may auto-start the unattended dev agent. See security.md.

Value Behavior
trusted (default) Auto-dispatch only issues whose author has write/maintain/admin access. Others get the warm reply + a maintainer control to dispatch by hand (human-in-the-loop against injected issue text).
all Auto-dispatch any author. Internal/trusted repos only.
never Never auto-dispatch; a maintainer always adds the try label.

triage.max_discussion_rounds

Field Default Meaning
max_discussion_rounds 4 Max back-and-forth rounds triage will hold with a reporter (each round = one triage reply) before it stops asking and loops in a human. The discussion loop fires on issue_comment from the reporter or a trusted maintainer (verdict discuss); this caps cost and prevents ping-pong. Requires the consumer's triage.yml wrapper to subscribe to issue_comment.

implement

Field Default Meaning
build_cmd — The build + test command the dev agent must get green before opening a PR (e.g. make build && make test, cargo test, npm test).
allow_network false If false, the implement/revise agent's harness denies direct network egress (curl/wget/ssh/…) so a prompt-injected issue can't exfiltrate. Set true only if your build genuinely needs the agent to reach the network. Not a full sandbox — see security.md.

testing

Optional pre-merge staging soak. Omit the block (or set enabled: false) to keep the original behavior: a review APPROVE auto-merges a simple PR and leaves everything else for a human. When enabled, an APPROVE-d PR is classified — a simple PR (within fast_path) keeps the auto-merge fast path; a complex PR is deployed to your testing environment, soaked, then labeled staging-soaked for a human to merge (Olympus never auto-merges a soaked PR). Olympus orchestrates the soak; your repo supplies the deploy + health commands. Requires the consumer pr-soak.yml wrapper. Applies to both agent and human PRs (gated by the same auto-merge trust). See cookbooks.

Field Default Meaning
enabled false Turn the soak gate on.
deploy_cmd — Run on the checked-out PR head to deploy it to the testing environment (e.g. make deploy-staging). $PR_NUMBER is available.
health_cmd — Polled repeatedly during the soak; non-zero exit = unhealthy. Omit to fall back to an HTTP 2xx check against observer.health_url.
soak_minutes 30 How long the PR must stay healthy before it's marked staging-soaked. Must be under the pr-soak workflow's timeout_minutes.
teardown_cmd — Optional; run after the soak (pass or fail) to tear the deployment down.
fast_path.max_loc triage.gates.max_loc A "simple" PR changes ≤ this many lines (additions + deletions).
fast_path.max_files triage.gates.max_files A "simple" PR changes ≤ this many files.
fast_path.areas — If set, a simple PR may touch only these path prefixes. Empty = areas not considered.

observer

Field Default Meaning
service_name the service Name used in incident text.
health_url — Health endpoint argus polls. (For the systemd path, set via the unit's ARGUS_HEALTH_URL instead.)
repo — (required to file) owner/name the incident issue is filed on.
labels incident Comma-separated issue labels.
readiness.jq — Optional jq filter over the health JSON for "parked" (up-but-not-working) detection. Omit → DOWN-only.
readiness.expect true Value the jq filter must equal; anything else = parked.

model

Field Default Meaning
model claude-3-5-sonnet-20241022 Model name sent to the gateway. Overridden by harness.model if set.

harness

Which agent CLI drives the surfaces. Omit the whole block for the built-in claude (Claude Code) harness — that's back-compatible with every existing consumer.

Field Default Meaning
harness.kind claude claude = built-in Claude Code. codex = built-in OpenAI Codex (codex exec). custom = run harness.command.
harness.command — custom only. A shell command template run per agent invocation. Placeholders: {model} {prompt_file} {out} {tools} {write} {max_turns}. e.g. aider --model {model} --message-file {prompt_file} > {out}.
harness.model — Model id for this harness; overrides .model. For codex/custom, the model the CLI is told to use (e.g. gpt-5).
harness.proxy — non-claude only. Egress proxy URL exported as HTTPS_PROXY/HTTP_PROXY/ALL_PROXY for the harness child, e.g. http://proxy.internal:8888. codex reaches its backend through it on staging/testing where a direct connection is blocked; the claude harness is never proxied (it uses the internal gateway). Prefer the HARNESS_PROXY repo secret to keep an internal IP out of committed config.
harness.health_probe true (false for codex) Poll the gateway's OpenAI-compatible /v1/models before each run (and to detect mid-run outages). Defaults off for codex (its backend isn't OpenAI-compatible); set explicitly to override.

For the built-in codex harness, put OPENAI_API_KEY on the runner (or as a repo secret) and set harness.proxy / the HARNESS_PROXY secret on staging/testing. See docs/cookbooks for the codex walkthrough + the prompt-shape constraint, and security.md for the egress note (codex lacks the claude harness's tool deny-list).

Example

See examples/consumer/.olympus.json for a complete, commented-by-structure config.