Skip to content
Back to skills

Workflow Review Deep

ASecurity

Orchestrate a multi-agent workflow to perform a deep-dive security, correctness, and architecture review.

  • 2 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added September 8, 2026
code-qualitycode-reviewsecurityperformance

Security analysis

A100/100

Pro scans all 5 files and shows the line behind each finding

Scanned September 8, 2026

npx -y skills add nobodyonlyc/skills --skill workflow-review-deep --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Workflow Review Deep?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Workflow Review Deep
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/nobodyonlyc-workflow-review-deep/badge)](https://www.skillsdirectory.com/skills/nobodyonlyc-workflow-review-deep)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: workflow-review-deep
description: Orchestrate a multi-agent workflow to perform a deep-dive security, correctness, and architecture review.
---

Deep multi-agent code review. PR number or branch: $ARGUMENTS

Multi-agent deep review: fan out one reviewer per changed file, then consolidate into a single verdict. Each phase has a detailed playbook in [`references/`](references/); see [`examples/`](examples/) for a full worked run.

## Skills this workflow drives
- [check-code-review](../check-code-review/SKILL.md) — correctness, performance, and simplification review per file (Phase 2).
- [check-security-review](../check-security-review/SKILL.md) — injection, secrets, auth, and input-validation review per file (Phase 2).
- [check-refactor](../check-refactor/SKILL.md) — shape the suggested cleanups when the user opts to apply fixes (Phase 3).
- [check-pr-review](../check-pr-review/SKILL.md) — frame the consolidated verdict as a PR review (Phase 3).

## Phases
1. **Gather & Fan-out** → [references/phase-1-gather-fanout.md](references/phase-1-gather-fanout.md)
   Collect the changed files (PR or branch diff) and spawn one review agent per file (batch beyond 8).
2. **Per-file Review** → [references/phase-2-per-file-review.md](references/phase-2-per-file-review.md)
   Each agent reviews its file for correctness, security, performance, and simplification.
3. **Consolidate & Verdict** → [references/phase-3-consolidate-verdict.md](references/phase-3-consolidate-verdict.md)
   Merge, dedupe, sort by severity, give a verdict, and optionally apply fixes.

## Hard gates
- Sort every finding by severity (🔴 Critical / 🟡 Important / 🔵 Suggestion) before presenting.
- Give one explicit verdict: **Approve** / **Request Changes** / **Needs Discussion**, and ask before applying any fix.

Files in this skill

  • SKILL.md1.8 KB
  • examples/example-deep-review.md1.1 KB
  • references/phase-1-gather-fanout.md572 B
  • references/phase-2-per-file-review.md956 B
  • references/phase-3-consolidate-verdict.md887 B

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…