ReviewGuard Adoption Plan
1. Why Use It Here
- You already have policy-driven engineering goals.
- Agent-led development benefits from deterministic PR gating.
- ReviewGuard helps enforce contract/security/governance checks early.
2. Rollout Strategy
- Week 1: run all rules in warn mode.
- Week 2: promote high-signal rules to error:
- contract-first
- hardcoded-secret block
- mutation-tests-required
- Week 3+: tune and promote remaining rules.
3. Source Documents for Policy Mapping
agent_policy.yaml
Agent_Enforcement.md
CI_Enforcement_Checklist.md
openapi.spectral.yaml
4. Minimum Blocking Rules (Day 1)
- Contract updates required when API behavior changes.
- Hardcoded secret patterns blocked.
- Backend mutation changes require tests.
- API spec changes require compatibility note.
5. Recommended Non-Blocking Rules Initially
- ADR reference checks.
- Broad authz heuristics (can be noisy until tuned).
6. Integration Notes
- Keep ReviewGuard as PR governance layer.
- Keep CI for deep validation:
- OpenAPI lint/breaking checks
- SAST/DAST/dependency scans
- integration/E2E tests
7. Success Criteria
- Reduced policy violations merged to main.
- Lower review burden for repeated rule classes.
- Stable false-positive rate after tuning.