# Code Review Orchestrator

> Triggered on pre-commit hook or when user says "code review", "review all", "quality check", "audit Use when this capability is needed.

- Skill: `tomevault-io/code-review-orchestrator-2` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add tomevault-io/code-review-orchestrator-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/code-review-orchestrator-2/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: tomevault-io (https://skillmd.com/u/tomevault-io)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tomevault-io/code-review-orchestrator-2

---


## Activation Contract

Triggered on pre-commit hook or when user says "code review", "review all", "quality check", "audit
code", "orchestrator", or similar.

## Hard Rules

- MUST run Security + Quality quick scan on every pre-commit
- MUST run all 7 dimensions pre-merge
- MUST block commit on CRITICAL findings (exit code 1)
- MUST generate a report at docs/code-reviews/
- MUST classify every finding by severity
- MUST NOT modify source files during review
- MUST keep quick scan under 30 seconds

## Decision Gates

| Gate            | Options                                                      | Rule                                   |
| --------------- | ------------------------------------------------------------ | -------------------------------------- |
| Scope           | all, security, quality, testing, docs, api, git, quick, full | quick = Security+Quality; full = all 7 |
| Severity action | CRITICAL block, HIGH warn+review, MEDIUM info, LOW suggest   | CRITICAL always blocks (exit 1)        |
| Trigger         | pre-commit auto vs manual                                    | Auto runs quick scope only             |

## Execution Steps

1. Detect trigger (pre-commit hook or manual request)
2. Resolve scope — quick (Security+Quality) or full (all 7)
3. Load config from configs/review-config.json
4. Execute scans for selected dimensions
5. Classify each finding by severity
6. Generate report (console + file at docs/code-reviews/)
7. Block if CRITICAL found (exit 1), otherwise allow (exit 0)

## Output Contract

Return severity breakdown, categorized issues (file:line, description, recommendation), report file
path, and exit code (1 if blocked).

## References

- `references/seven-dimensions.md` — Dimension descriptions and scope mapping
- `references/severity-matrix.md` — Severity levels with actions
- `references/commands.md` — Full command reference and performance
- `references/report-format.md` — Console and file report templates
- `references/judgment-day.md` — Dual review protocol
- `references/config-example.md` — Review configuration JSON
- `references/integration.md` — Git hooks and CI/CD setup

---
> Source: [EmmanuelOrtiz87/gentle-vanguard](https://github.com/EmmanuelOrtiz87/gentle-vanguard) — distributed by [TomeVault](https://tomevault.io).
<!-- tomevault:4.0:skill_md:2026-05-22 -->

