weitzer-org/GSR Code Review
Multi-agent Gemini-powered code review that posts inline findings on a pull request.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| gemini-api-key | Gemini API key used to run the review. Store as a repo/org secret. | yes | — |
| github-token | Token used to read the PR diff and post review comments. | no | ${{ github.token }} |
| mode | Review strategy: "subagent" runs the full swarm of specialized agents (Architecture, Logic, Security, TechDebt, Testing, ...) plus a deduplication pass and finds more issues but costs more tokens/time. "basic" runs a single general-purpose review pass, faster and cheaper. | no | subagent |
| fail-on-severity | Fail the workflow if a finding at or above this severity is reported: one of "critical", "high", "medium", "low", or "none" to never fail. | no | none |
| gemini-model | Override the Gemini model used for review. | no | "" |
| max-review-files | Maximum number of changed files to review before truncating. | no | 300 |
| agents | Comma-separated list of subagent IDs to run (e.g. "logic,security"), or "all" to run every agent. Only applies when mode is "subagent"; ignored (with a warning) when mode is "basic". Defaults to "all". | no | all |
Outputs
no outputs