wuisabel-gif/SecondOpinion Agent-Agnostic PR Review
Agent-agnostic AI pull request reviews with interchangeable model backends
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| api-key | API key for the selected review provider; optional for unauthenticated local services | no | "" |
| provider | anthropic, openai-responses, openai-compatible, or webhook | no | anthropic |
| model | Provider-specific model override | no | "" |
| base-url | Base URL for OpenAI-compatible APIs | no | "" |
| endpoint | Exact endpoint URL for the generic webhook adapter | no | "" |
| auth-header | Authentication header for compatible APIs and webhooks | no | Authorization |
| auth-scheme | Authentication scheme, such as Bearer or none for a raw key | no | Bearer |
| response-json-pointer | Optional JSON Pointer locating a normalized review in a webhook response | no | "" |
| passes | Number of independent model passes, from 1 to 7 | no | 1 |
| vote-threshold | Votes required to keep a finding; defaults to a strict majority | no | "" |
| context-bytes | Maximum repository-context bytes; set to 0 to disable context | no | 60000 |
| bot-login | GitHub login whose prior comments are trusted for deduplication | no | github-actions[bot] |
Outputs
no outputs