jaibhasin/PRBot Precision Review
Precision-first PR reviewer powered by OpenRouter, runs as a GitHub Action.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| openrouter_api_key | OpenRouter API key used for LLM review calls. | no | "" |
| github_token | GitHub token with pull request write and checks write permissions. If omitted, the workflow job token (GITHUB_TOKEN) is used. | no | "" |
| pr_number | Pull request number. When empty, PRBot reads it from the GitHub event payload. | no | "" |
| dry_run | Build and print the contextual review manifest without calling the LLM or posting comments. | no | false |
| review_model | OpenRouter model used by the primary review agent. | no | deepseek/deepseek-v4-flash |
| verification_model | Independent OpenRouter model used to verify candidate findings. | no | deepseek/deepseek-v4-flash |
| max_review_minutes | Hard wall-clock review deadline. | no | 15 |
| max_input_tokens | Maximum estimated input tokens across all model calls. | no | 500000 |
| max_cost_usd | Maximum estimated model cost in US dollars. | no | 3.00 |
| max_concurrency | Maximum number of concurrent model calls. | no | 8 |
| max_comments | Maximum number of verified inline findings published per run. | no | 12 |
| engine | Review engine: contextual (default) or legacy rollback. | no | contextual |
Outputs
no outputs