# Coordinate Multi-Agent Code Reviews

Single-perspective reviews can miss security, architecture, performance, and quality concerns. This skill coordinates specialized reviewers and synthesizes their findings into one prioritized assessment.

## Install

```bash
npx skillstore add sickn33/error-debugging-multi-agent-review
```

## Metadata

- Status: approved
- Slug: sickn33-error-debugging-multi-agent-review
- Skillstore revision: r2
- Version status: missing
- Tree hash: e9880e8b1dee332056d6fee10e3821ba50b8b8059ceedda869b9e0ff721f5352
- Author: sickn33
- GitHub username: sickn33
- License: MIT
- Repository: https://github.com/sickn33/antigravity-awesome-skills/tree/main/skills/error-debugging-multi-agent-review
- Ref: 88a8e9a07f4c54ab105c1c41b6267c287146b07b
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands
- Quality score: 78
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/sickn33-error-debugging-multi-agent-review
- Manifest: https://skillstore.pages.dev/api/skills/sickn33-error-debugging-multi-agent-review/manifest

## Capabilities

- Routes review targets to specialized agent roles based on context.
- Supports parallel, sequential, and hybrid review workflows.
- Shares review context and intermediate insights across agents.
- Consolidates findings into critical issues, important issues, and improvements.
- Resolves conflicting recommendations through weighted review logic.
- Applies quality validation and confidence scoring to synthesized results.

## Use Cases

- Review a Feature Before Merge: Coordinate code quality and security perspectives, then combine duplicate findings into a prioritized merge assessment.
- Assess Application Risk: Route application code through security, compliance, and architecture reviews while preserving evidence and confidence levels.
- Evaluate Release Readiness: Combine architecture, performance, testing, and deployment reviews into a sequenced decision with clear blockers.

## Prompt Templates

### Basic Two-Perspective Review

```
Review [target] with code quality and security perspectives. List findings by severity, cite evidence, and state verification steps.
```

### Parallel Specialist Review

```
Coordinate parallel security, architecture, and performance reviews for [target]. Merge duplicate findings and explain disagreements before prioritizing actions.
```

### Sequential Release Review

```
Design a sequential review for [target]. Let architecture findings inform implementation review, then evaluate tests and deployment readiness.
```

### Advanced Hybrid Orchestration

```
Create a hybrid multi-agent review for [target] under [constraints]. Define routing, shared context, conflict resolution, confidence scoring, and final acceptance criteria.
```

## Limitations

- Provides orchestration guidance and illustrative pseudocode, not an executable review engine.
- Depends on available agent tools and the quality of each reviewer.
- Does not replace project-specific tests, runtime validation, or expert approval.
- References resources/implementation-playbook.md, but that file is not included.

## Best Practices

- Define review scope, constraints, and acceptance criteria before selecting agents.
- Keep independent reviews parallel, but sequence reviews that depend on earlier findings.
- Require evidence, confidence, and verification steps for every prioritized finding.

## Anti Patterns

- Do not assign every reviewer when the target only needs focused analysis.
- Do not merge conflicting recommendations without recording the disagreement and rationale.
- Do not treat synthesized advice as proof without tests or environment-specific validation.

## Security Audit

- Audited at: 2026-07-23T23:43:28.763\+00:00
- Summary: All 21 external-command findings are false positives caused by Markdown backticks or Python code fences; the examples contain no shell execution. The two reconnaissance findings are review-strategy headings, not host discovery, and no prompt injection or other semantic threat was found.

## Stats

- Views: 104
- Downloads: 14
- Favorites: 0
- Popularity score: 0