Skip to content

autoreview: default review effort to high, escalate to xhigh for high-risk diffs #24

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
justin808 wants to merge 1 commit into
base: main
Choose a base branch
from
+14 −6
Open

autoreview: default review effort to high, escalate to xhigh for high-risk diffs #24

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 14 additions & 6 deletions skills/autoreview/SKILL.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -105,16 +105,23 @@ review tooling described in the intro. If neither engine exists in the current e
tell the user which review engines are missing instead of improvising a different review scope. Pick
the command that matches Step 1:

**Reasoning effort.** Run review at `high` effort by default — it is the cost-effective depth for
routine diffs and never stalls the loop. Escalate to `xhigh` only for the high-risk categories in
the second-pass section below. Pass effort with `-c model_reasoning_effort=...` (the key mirrors
`config.toml`); override the per-environment default by exporting `CODEX_REVIEW_EFFORT`.

```bash
effort="${CODEX_REVIEW_EFFORT:-high}" # set to xhigh for high-risk diffs (see "High-risk second pass")

# Dirty local patch, including staged, unstaged, and untracked files.
codex review --uncommitted
codex review -c model_reasoning_effort="$effort" --uncommitted

# Branch or PR diff.
base=$(gh pr view --json baseRefName --jq .baseRefName 2>/dev/null || echo main)
codex review --base "origin/$base"
codex review -c model_reasoning_effort="$effort" --base "origin/$base"

# Single commit.
codex review --commit <sha>
codex review -c model_reasoning_effort="$effort" --commit <sha>
```

Prefer explicit target selection over custom focus text. Some Codex CLI versions reject a custom
Expand All @@ -125,7 +132,7 @@ When the installed CLI accepts focus text with the selected target flag, keep th
Step 1 and append the prompt there:

```bash
codex review --base "origin/$base" "Focus on performance- or framework-sensitive regressions (per AGENTS.md), generated output, and repo workflow correctness."
codex review -c model_reasoning_effort="$effort" --base "origin/$base" "Focus on performance- or framework-sensitive regressions (per AGENTS.md), generated output, and repo workflow correctness."
```

For longer instructions, create an ignored scratch file, for example
Expand All @@ -134,7 +141,7 @@ path. Read from stdin only when the selected review engine supports that mode wi
target:

```bash
codex review --base "origin/$base" - < .context/autoreview-focus.md # create this ignored scratch file first
codex review -c model_reasoning_effort="$effort" --base "origin/$base" - < .context/autoreview-focus.md # create this ignored scratch file first
```

Never silently switch the engine the user asked for. If the requested engine hits model
Expand All @@ -144,7 +151,8 @@ capacity, retry the same engine a few times rather than swapping it.

For high-risk changes in the `AGENTS.md` hosted-CI-ready, force-full hosted-CI, or benchmark
categories (labels per `AGENTS.md` → **Agent Workflow Configuration**), or when the user asks
for a panel/second model, run one additional review after the primary review is clean:
for a panel/second model: first set `effort="xhigh"` in Step 3 so the **primary** review runs at
full depth, then run one additional review after the primary review is clean:

- If the primary review used `codex review`, use Claude review tooling if it is available in the
current environment, such as `/code-review` or `/code-review ultra`.
Expand Down
Loading