Claude-code-agents full-audit

Run all 11 audit agents in parallel, then consolidate findings with fix-planner

install
source · Clone the upstream repo
git clone https://github.com/undeadlist/claude-code-agents
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/undeadlist/claude-code-agents "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/full-audit" ~/.claude/skills/undeadlist-claude-code-agents-full-audit && rm -rf "$T"
manifest: skills/full-audit/SKILL.md
source content

Full Audit Workflow

Run a comprehensive audit of the entire codebase. Use before major releases or as a weekly health check.

Phase 1: Parallel Audit

Spawn ALL of the following agents in parallel (single response with multiple Task calls):

AgentOutput FileFocus
code-auditor
.claude/audits/AUDIT_CODE.md
Code quality
bug-auditor
.claude/audits/AUDIT_BUGS.md
Runtime bugs
security-auditor
.claude/audits/AUDIT_SECURITY.md
OWASP deep scan
doc-auditor
.claude/audits/AUDIT_DOCS.md
Documentation gaps
infra-auditor
.claude/audits/AUDIT_INFRA.md
Infrastructure config
ui-auditor
.claude/audits/AUDIT_UI_UX.md
UI/UX and accessibility
db-auditor
.claude/audits/AUDIT_DB.md
Database performance
perf-auditor
.claude/audits/AUDIT_PERF.md
Performance issues
dep-auditor
.claude/audits/AUDIT_DEPS.md
Dependency health
seo-auditor
.claude/audits/AUDIT_SEO.md
SEO optimization
api-tester
.claude/audits/API_TEST_REPORT.md
API validation

Each agent targets the project's source code. Provide each with a prompt like:

Audit [target directory]. Save report to [output file].

Phase 2: Consolidation

After all auditors complete, spawn:

  • fix-planner
    - Read all
    .claude/audits/AUDIT_*.md
    files and create
    .claude/audits/FIXES.md
    with prioritized action items

Post-Audit Actions

  1. Review
    FIXES.md
    - Prioritize based on your timeline
  2. Run
    code-fixer
    - Implement P1 (critical) fixes
  3. Run
    test-runner
    - Verify fixes work
  4. Re-audit affected areas if needed