odysseustech/Auto-Resolve Bot Review Threads
Resolves addressed bot review comments on PRs using a deterministic gate chain and Claude Haiku.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| anthropic-api-key | Anthropic API key for the Haiku classifier. | yes | — |
| github-token | Token with pull-requests:write and contents:read. | no | ${{ github.token }} |
| dry-run | If true, log decisions but do not resolve threads. | no | false |
| bot-logins | Comma-separated allowlist of bot login names (without [bot] suffix is also accepted). | no | coderabbitai[bot],greptile-apps[bot],copilot-pull-request-reviewer[bot],cursor[bot] |
| model | Anthropic model id. | no | claude-haiku-4-5-20251001 |
| max-resolutions-per-run | Hard cap on auto-resolutions in a single run. | no | 20 |
| generated-file-globs | Comma-separated globs for generated files (auto-resolved without AI). | no | **/generated/**,**/*.generated.*,**/dist/**,**/build/**,**/package-lock.json,**/pnpm-lock.yaml,**/bun.lockb,**/yarn.lock,**/prisma/migrations/** |
| pr-number | PR number to process. Required when triggered via workflow_dispatch. | no | "" |
Outputs
| name | description |
|---|---|
| resolved-count | Number of threads resolved this run. |
| skipped-count | Number of threads skipped. |
| classified-count | Number of threads that reached the AI classifier. |