canblmz1/RuleProbe
Test whether your CLAUDE.md / AGENTS.md / .cursor/rules are actually followed at runtime
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| dir | Directory containing instruction files (CLAUDE.md, AGENTS.md, .cursor/rules). Defaults to repo root. | no | . |
| provider | AI provider to run scenarios against (mock, dry-run, gemini, openrouter, claude-code) | no | mock |
| extractor | Rule extraction mode (deterministic, ai-assisted, hybrid) | no | deterministic |
| fail-below | Fail the step if compliance score drops below this value (0-100). Set to 0 to never fail. | no | 0 |
| model | Model name for providers that support it (e.g. gemini-2.5-flash, openai/gpt-4o) | no | "" |
| version | ruleprobe-ai npm version to use (e.g. latest, 0.4.0) | no | latest |
| args | Extra CLI flags passed verbatim to ruleprobe run (e.g. --no-cache --debug-extractor) | no | "" |
| comment | Post compliance score as a PR comment (requires pull_request event and write permissions) | no | false |
Outputs
| name | description |
|---|---|
| score | Compliance score (0-100) |
| passed | Number of rules that passed |
| failed | Number of rules that failed |
| skipped | Number of rules that were skipped |
| report-path | Path to the generated JSON report |