Skip to main content
skill CX Codex

codebase-review-swarm · zaxbyhub, claude — Claude Code skill: how it works and setup

codebase-review-swarm executes quote-grounded codebase audits across security, QA, and performance. Suitable for full-repository assessments where no source files may be touched.

MIT ★ 539 updated 2026-10-29

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add ZaxbyHub/opencode-swarm --skill codebase-review-swarm -a claude-code
Open on GitHub

What codebase-review-swarm does

This workflow runs deep review sweeps across an entire repository or major subsystems. It is strictly read-only: the agent never edits source files. All generated findings and logs land only inside .swarm/review-v8/runs/<run_id>/.

Execution starts with a Phase 0 inventory. The process halts for track selection unless the user already specified target audit areas. The skill enforces a strict evidence policy: every claim requires an exact code quote. Findings pass through separate reviewer and critic stages before the summary report is finished.

This setup will fail you if you need bug fixing or code changes, because it exclusively produces diagnostic reports.

What it helps with

  • audit large codebases without editing any production files
  • produce evidence-backed review reports grounded in exact source quotes
  • scan repositories for security vulnerabilities, QA gaps, and AI slop
  • run multi-stage reviewer and critic passes on architecture findings

Example prompts for the agent

  • Run a quote-grounded security and QA audit across the repository using codebase-review-swarm
  • Execute Phase 0 inventory and wait for track selection before auditing the auth module
  • Inspect the frontend codebase for accessibility defects and AI slop

How to install codebase-review-swarm

1Skills CLI (Vercel)

npx skills add ZaxbyHub/opencode-swarm --skill codebase-review-swarm -a claude-code

2Manually: put the files in place

~/.claude/skills/codebase-review-swarm/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/codebase-review-swarm/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about codebase-review-swarm

How to install codebase-review-swarm in Claude Code?

Run npx skills add ZaxbyHub/opencode-swarm --skill codebase-review-swarm -a claude-code. The workflow will be saved to ~/.claude/skills/codebase-review-swarm/SKILL.md.

What does codebase-review-swarm do?

It coordinates a thorough read-only repository audit across security, performance, QA, and accessibility. Every claim in the generated report must include direct code quotes.

Does codebase-review-swarm modify any source code?

No. The skill operates in read-only mode and writes artifacts exclusively under .swarm/review-v8/runs/<run_id>/.

Source text

SKILL.md · MIT
---
name: codebase-review-swarm
description: >
  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.
metadata:
  audience: swarm-plugin
---

# 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.

© ZaxbyHub · MIT · ZaxbyHub/opencode-swarm

Similar skills

All skills for Code review and refactoring →

codebase-review-swarm by other authors comes first.

codebase-review-swarm ZaxbyHub codebase-review-swarm runs quote-grounded code reviews and security or QA audits without editing project files. Built for deep inspections across entire codebases or large subsystems. CX Codex +1 ★ 539repo stars codebase-review-swarm ZaxbyHub The codebase-review-swarm skill audits a repository and ties each finding to source evidence. Use it for focused security, QA, accessibility, performance, or other review tracks. CX Codex +1 ★ 539repo stars review-swarm Dimillian review-swarm runs a parallel read-only review of a git diff or selected files. Use it to surface regressions, security concerns, and missing test coverage before merging. CX Codex +1 820installs review-loop sangrokjung review-loop runs an adversarial verification loop where a fresh, isolated checker inspects behavioral changes until reaching a strict APPROVE verdict. CX Codex +1 ★ 880repo stars adversarial-review syahiidkamil adversarial-review subjects fresh builds to hostile inspection using isolated reviewer contexts and mandates empirical reproduction on the running system. CX Codex +1 ★ 406repo stars code-review-swarm ruvnet code-review-swarm deploys coordinated AI review agents against pull requests via the GitHub CLI. It flags security flaws and architecture drift before merging. CX Codex ★ 850repo stars codebase-standards devvmhq The codebase-standards Claude Code skill checks devvm changes against project standards. Run it before a commit or while reviewing a branch or change. CX Codex +1 ★ 738repo stars ac-reviewing-codebase souliane ac-reviewing-codebase — a skill for holistic review of a repo portfolio: it checks skill quality, code health and infrastructure alignment across repositories. CX Codex +1 1,000installs

From the same set121

The whole set →
gate-attribution gate-attribution sets the rule that keeps reviewer and test_engineer checks attributable: one taskId per dispatch, with lanes run in parallel to save overhead. CX Codex +1 ★ 539repo stars swarm-implement Adapts the canonical opencode-swarm implementation workflow for multi-file features, refactoring, and structured validation. CX Codex +1 ★ 539repo stars opt-in-tool-registration The opt-in-tool-registration skill sets a pattern for registering opt-in tools through a separate map and conditional merging in agents/index.ts. Use it when adding a feature controlled by a flag. CX Codex +1 ★ 539repo stars ensure-all-tests-pass-after-code-changes ensure-all-tests-pass-after-code-changes — a skill that keeps the agent honest about tests: run them after an edit, and don't hand a failing branch to review. Test updates ship with the code. CX Codex +1 ★ 539repo stars add-path-type-assertions-or-realpath-resolve-calls-in-path-joini add-path-type-assertions-or-realpath-resolve-calls-in-path-joini is a Claude Code skill that adds path-type checks or realpath/resolve calls to path-joining utilities. It activates around path.join, realpath, and fs path work. CX Codex +1 ★ 539repo stars run-failing-tests-immediately-after-test-gate-failure run-failing-tests-immediately-after-test-gate-failure is a Claude Code skill that makes the agent rerun failing tests when test_gate fails and repair the root cause before claiming the task is done. CX Codex +1 ★ 539repo stars run-tests-on-the-unmodified-codebase-before-starting-to-establis run-tests-on-the-unmodified-codebase-before-starting-to-establis is a Claude Code skill that requires running tests on untouched code before edits, recording failures to distinguish them from new regressions. CX Codex +1 ★ 539repo stars engineering-conventions The engineering-conventions skill sets a workflow for changes that could affect opencode-swarm’s engineering invariants. Use it when editing architecture, tests, or session state. CX Codex +1 ★ 539repo stars

Project sponsors

Companies that keep this analytics open