enriconunes/Claude AutoFix
Automatically analyses and fixes failing tests using Claude AI (Python/pytest, JS/TS with Vitest or Jest)
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Activelast commit Apr 20, 2026
- License
- MIT
Pinned Snippet
uses: enriconunes/claude-autofix-action@86cc82d4d18e33e99106982f12b5e2a502d5ab97 # v2.1.2tags can be moved; commit SHAs can't. why a SHA?
Inputs
| name | description | required | default |
|---|---|---|---|
| anthropic-api-key | Anthropic API key (ANTHROPIC_API_KEY) | yes | — |
| max-fixes | Maximum number of fixes per run | no | 5 |
| python-version | Python version to use | no | 3.11 |
| create-pr | Automatically create a Pull Request with the fixes (true/false) | no | true |
| test-framework | Test framework: pytest | vitest | jest | no | pytest |
| language | Source language: python | typescript | javascript | no | python |
| enable-error-log | Append failing test info to logs/error_history.jsonl in main (true/false). Does not consume Claude tokens. | no | false |
| enable-insights | Post a Claude-powered insight comment on the PR based on the author's error history (true/false). Requires enable-error-log on previous runs to have data. Consumes Claude tokens. | no | false |
Outputs
| name | description |
|---|---|
| fixes-applied | Number of fixes successfully applied |
| pr-url | URL of the Pull Request created with the fixes |