Audit Fix Agent — Security agent for Claude Code
Applies a single verified audit finding as a code fix in one file.
How to install Audit Fix Agent
Installs to ~/.claude/agents/raaaf-claude-skills-audit-fix-agent.md
mkdir -p ~/.claude/agents && curl -fsSL https://raw.githubusercontent.com/raaaf/claude-skills/HEAD/agents/audit-fix-agent.md -o ~/.claude/agents/raaaf-claude-skills-audit-fix-agent.md Restart Claude Code, or start a new session, for it to be picked up.
What Audit Fix Agent does
name: audit-fix-agent description: Applies a single verified audit finding as a code fix in one file. Runs in parallel with sibling fix agents in a shared working tree. Used by the /audit and /full-audit skills' Step E, never dispatched directly by the user. tools:
- Read
- Edit
- Write
- MultiEdit
- Grep
- Glob
- Bash model: sonnet effort: medium
Skill-frontmatter hooks apply to the main session only, not to subagents
(docs: hooks.md "Hook scope"). Without this block the w
Alternatives in Security
- Offensive Security Engineer — Red-team the verification engine 179 ★
- Scope Audit — SCOPE audit orchestrator — single entry point for the full audit pipeline 54 ★
- Duplication Hunter — Hunts code duplication, redundant utilities, and parallel implementations across an entire codebase 42 ★
Full documentation available on GitHub
View Source RepositoryRelated Agents
Audit Fix Verifier
Verifies a fix applied by audit-fix-agent actually resolves the original finding and introduces no regression.
Narwhal Duplication
Duplicate-content & canonicalization specialist for the narwhal audit — near-duplicate page clusters and wheth
Worthiness Evaluator
Advisory test-worthiness audit for the emergent surface. Reads only the phase''s changed test files plus their
ForensicAgent
Forensic security analyst — PII detection, secret scanning, identity leak auditing across git history, staged
Fallow
Run a codebase-intelligence audit (fallow) scoped to working-tree changes and report raw findings — dead code,
P Code Review Security
Invoked by p-code-review to run the security channel across the entire working tree (no diff). Adversarial rea
Related Skills
UX Fix
Opt-in fix loop. Reads the latest report from .ux/, validates working tree, and applies findings as atomic com
Sibling Sweep
Sibling sweep — find the members the fix never reached
Zp Report
Draft a submission-ready report from a verified finding. Never submits. Usage: /zp-report