# Codebase Review Swarm

> Claude Code adapter for running a rigorous, quote-grounded codebase review or security/QA/accessibility/performance/AI-slop/enhancement audit. Use for full-repo or large-subsystem review reports; not for normal implementation.

- Skill: `zaxbysauce/codebase-review-swarm-2` (Agent Skill)
- Install (CLI): `npx skillmds@latest add zaxbysauce/codebase-review-swarm-2`
- Raw SKILL.md: https://api.skillmd.com/api/skills/zaxbysauce/codebase-review-swarm-2/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Web & Frontend
- Author: zaxbysauce (https://skillmd.com/u/zaxbysauce)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/zaxbysauce/codebase-review-swarm-2

---


# Codebase Review Swarm

Read and follow `../../../.opencode/skills/codebase-review-swarm/SKILL.md` as
the canonical workflow.

## Claude Code Execution Notes

- Treat this as a read-only review workflow. Do not modify source files.
- Write review artifacts only under `.swarm/review-v8/runs/<run_id>/`.
- Run Phase 0 inventory first and stop for review-mode selection unless the
  user already selected tracks and explicitly authorized continuing.
- Preserve the skill's evidence rule: no quote, no claim.
- For dispatch lanes, only if the session actually exposes the controller's
  lane tools, treat `output` as a preview and call `retrieve_lane_output` for full `output_ref` artifacts
  before consuming ledgers, counting candidates, routing validation, or
  declaring coverage clean; when those tools are absent, treat inline
  `output` as the full artifact and disclose any truncation as an explicit
  coverage limitation.
- Use independent reviewer and critic passes before finalizing report findings.