/pre-push-review¶
Review class · gate
Advise only. Reads changes, reports findings, and never writes a file or commits. Shared rules: CONVENTIONS.md.
Runs up to five review dimensions (secrets, correctness, tests/build, dependencies, style) via a Workflow-tool pipeline — the secret scan first as a gate, then the rest in parallel — and aggregates one PASS / WARN / BLOCK verdict with a report artifact. Cost-aware: it states the scope and agent count and asks before spawning, skips dimensions with nothing to review, and on a re-run defaults to only the dimensions that still have findings. Supports a non-interactive CI mode with an exit-code contract.
At a glance¶
| Run it | /pre-push-review |
| Class | review · gate |
| Version | 0.5.0 |
| Author | navjyotnishant |
| Cost | up to 5 dimension agents in parallel after the secret-scan gate, plus 1 report-writer after aggregation (Step 5.5) — so up to 6 |
| Needs | gitleaks or trufflehog or detect-secrets · the repo's own test/lint/build commands |
| Source | skills/pre-push-review/SKILL.md |
What it needs, and what happens without it¶
Every tool is detected at runtime — none is installed for you.
| Tool | Without it |
|---|---|
gitleaks or trufflehog or detect-secrets |
BLOCK — the gate cannot be skipped |
| the repo's own test/lint/build commands | SKIP that dimension and say so |
The pipeline¶
Agents it spawns¶
correctness-reviewer— Review a code diff for correctness bugs — logic errors, regressions, unhandled edge cases, and missing…dependency-reviewer— Review dependency-manifest and license changes in a code diff — added/removed/upgraded packages, license…review-report-writer— Turn an aggregated pre-push review result into a self-contained, human-readable HTML report — verdict banner…secrets-reviewer— Use this agent for a deeper semantic security review of a code diff AFTER a local secret scan has already…style-reviewer— Review a code diff for style, consistency, and hygiene — does new code match the conventions of the code…tests-build-runner— Auto-detect and run a repo's own test/lint/build commands over the current changes and triage any failures.…
The procedure¶
The executable steps live in the skill file itself and are deliberately not reproduced here: they are instructions to the model at runtime, not documentation.