anthropics / anthropics/claude-code-action
Auto-review should post a comment when no issues are found (silent clean pass)
- Dominant language
- TypeScript
- Stars
- 8.9k
- Forks
- 2.1k
- Avg merge
- 3d 9h
- Merged PRs (30d)
- 10
Description
**Problem**
When `claude-code-action` runs an automatic PR review (triggered by `pull_request` event with a `prompt`) and finds no issues, it exits silently — no PR comment, no inline comments, no indication that it ran and reviewed the code.
The action runs successfully (5+ minutes, exit code 0, "No buffered inline comments" in logs), but the PR author has no way to know whether:
1. The review ran and found nothing (clean pass), or
2. The review silently failed or didn't actually analyse the code
**Expected behaviour**
Post a brief PR comment confirming the review completed with no findings, e.g.:
> ✅ Claude Code reviewed this PR and found no issues.
This is what Gemini Code Assist does — when it has no findings, it explicitly says so. This is much more helpful than silence, because it gives the author confidence that the review actually ran.
**Context**
- Using `anthropics/claude-code-action@v1` with `prompt` input (not `@claude` trigger)
- Trigger: `pull_request: [opened, synchronize, ready_for_review, reopened]`
- The `@claude review` comment-triggered path works correctly and posts detailed reviews
- The automatic path runs but produces no visible output on clean passes
- `show_full_output: false` (default), so there's no way to see what Claude thought
**Workaround**
None currently — we rely on the `@claude review` comment trigger for visible feedback and treat the auto-review as a silent safety net, which reduces its value.
**Suggestion**
Add an option (or make it default) to post a brief "no issues found" comment when the review completes without findings. Even a one-line sticky comment would suffice.
Contributor guide
Research direction
Start by tracing the automatic pull_request review path described in the issue, especially the handling of a successful run with no buffered inline comments. Confirm how the action currently decides whether to publish output, then verify that a clean pass produces the requested PR comment without changing the existing comment-triggered review path.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- github-actions, typescript
- Domain
- ci-cd, devtools
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 48/100