architcybercrime/AI Code Reviewer
Review pull requests with Claude — inline comments, summary, incremental + dedup.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Activelast commit May 30, 2026
- License
- MIT
Pinned Snippet
uses: architcybercrime/ai-code-reviewer@acdee5915c79f04d38e023890590a2f6aa4d3123 # v0.1.0tags can be moved; commit SHAs can't. why a SHA?
Inputs
| name | description | required | default |
|---|---|---|---|
| anthropic_api_key | Anthropic API key for the Claude review model. | yes | — |
| github_token | GitHub token with pull-requests: write. | no | ${{ github.token }} |
| model | Primary review model. | no | claude-opus-4-7 |
| summary_model | Model used for the PR-level summary. | no | claude-haiku-4-5 |
| effort | Thinking effort for the review model: low | medium | high | max. | no | high |
| max_files | Maximum number of files to review (<=0 means no limit). | no | 60 |
| min_severity | Only post findings at or above this severity: '' | critical | major | minor | suggestion. | no | "" |
| post_check_run | Also post a GitHub Check Run with findings as annotations (needs checks: write). | no | false |
| path_filters | Newline-separated glob include/exclude rules (use ! to exclude). | no | "" |
| dry_run | Run the review but do not post comments to GitHub. | no | false |
Outputs
| name | description |
|---|---|
| findings | Number of findings posted. |
| files_reviewed | Number of files reviewed. |
| skipped | Number of files skipped. |