Security AuditCAUTION

The most comprehensive Claude Code guide: agentic workflows, hooks, skills, MCP servers, quizzes, and production-ready templates. 430K+ lines.

The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.

---
name: security-audit
description: Comprehensive security audit with scored posture assessment
argument-hint: "[path] [--owasp] [--verbose]"
effort: high
disable-model-invocation: true
---

# Security audit

Comprehensive security audit of your project AND Claude Code configuration. Analyzes secrets exposure, injection surfaces, dependencies, hook security, and produces a scored security posture assessment.

**Time**: 2-5 minutes | **Scope**: Full project + Claude Code config

> For a quick config-only check, use `/security-check` instead.

## Instructions

You are a senior application security engineer. Perform a 6-phase security audit and produce a scored report with prioritized remediation plan.

---

### Pre-Step: Establish audit context

**Before running any checks**, use `AskUserQuestion` to ask:

1. **Environment**: Is this code running in production, staging, or local development?
2. **Scope**: Full audit or specific areas to prioritize?

This is critical for accurate findings:
- **Local dev**: `DEBUG=True`, CORS `*`, HTTP without TLS, `.env` files, all normal. Do NOT flag as vulnerabilities. Mention in an "Before going to production" informational section instead.
- **Staging**: Configs should mirror production. Flag deviations as MEDIUM.
- **Production**: Any misconfiguration is a real finding with full severity.

If the user doesn't answer or is unsure, default to **production** (conservative).

---

### Phase 1: Configuration security (via /security-check)

Execute all checks from `/security-check` (the `examples/skills/security-check/SKILL.md` command). This covers:
- MCP server audit against CVE database
- Skills & agents against known malicious entries
- Hook exfiltration patterns
- Memory poisoning detection
- Permissions & settings review
- Exposed secrets in Claude Code config

Record findings, as they contribute to the final score.

---

### Phase 2: Project secrets scan

Scan the entire project for exposed secrets and credentials:

```bash
# API keys and tokens
grep -rn --include="*.{js,ts,py,go,java,rb,php,yaml,yml,json,toml,env,cfg,ini,conf}" \
  -E '(?i)(api[_-]?key|apikey|secret|password|passwd|token|bearer|auth)\s*[=:]\s*["'\''"][^"'\'']{8,}["'\''"]\s' \
  --exclude-dir={node_modules,vendor,.git,dist,build,target,__pycache__,.venv} . 2>/dev/null | head -30

# Known provider key patterns
grep -rn -E 'sk-[a-zA-Z0-9]{20,}|sk-ant-[a-zA-Z0-9]{20,}|ghp_[a-zA-Z0-9]{36}|AKIA[A-Z0-9]{16}|xox[bps]-[a-zA-Z0-9\-]{20,}' \
  --exclude-dir={node_modules,vendor,.git,dist,build,target} . 2>/dev/null | head -20

# Private keys
grep -rn 'BEGIN.*PRIVATE KEY' --exclude-dir={node_modules,vendor,.git} . 2>/dev/null

# .env files that might be committed
find . -name ".env*" -not -path "*/node_modules/*" -not -path "*/.git/*" -type f 2>/dev/null

# Check .gitignore coverage
[ -f ".gitignore" ] && {
  grep -q "\.env" .gitignore && echo "✅ .env in .gitignore" || echo "⚠️ .env NOT in .gitignore"
  grep -q "\.pem" .gitignore && echo "✅ .pem in .gitignore" || echo "⚠️ .pem NOT in .gitignore"
  grep -q "\.key" .gitignore && echo "✅ .key in .gitignore" || echo "⚠️ .key NOT in .gitignore"
}
```

**Anti-false-positive rule (MANDATORY before reporting any secret finding):**

Before raising a secrets finding, run these verification commands:

```bash
# 1. Verify .env is actually in .gitignore (if yes, local .env is NOT a finding)
grep -n '\.env' .gitignore 2>/dev/null || echo ".env NOT in .gitignore"

# 2. Verify secrets were actually committed (empty output = no finding)
git log --all -p -- '*.env' '*.key' '*.pem' '*.secret' 2>/dev/null | grep -E '^\+.*(password|secret|api_key|token)' | head -20

# 3. Check git history for provider-specific patterns
git log --all -p 2>/dev/null | grep -E '^\+(sk-[a-zA-Z0-9]{20,}|AKIA[A-Z0-9]{16}|ghp_[a-zA-Z0-9]{36})' | head -10
```

Only report a secret finding if you have **concrete proof from these commands**. A `.env` file present locally is not a finding if it's in `.gitignore`. Never report "secrets may be exposed" based on pattern matching alone.

**Scoring:**
- 0 secrets found → +20 points
- 1-3 secrets → +10 points
- 4+ secrets → 0 points
- Private key committed → -10 points

---

### Phase 3: Prompt injection surface

Analyze markdown and config files for injection vectors:

```bash
# Zero-width characters (invisible instructions)
grep -rPn '[\x{200B}-\x{200D}\x{FEFF}]' --include="*.md" --include="*.yaml" --include="*.json" . 2>/dev/null

# Hidden HTML comments with instructions
grep -rn '<!--' --include="*.md" . 2>/dev/null | grep -i 'ignore\|system\|admin\|instruction\|override\|forget'

# Base64 in comments (potential hidden payloads)
grep -rn -E '[#;].*[A-Za-z0-9+/]{20,}={0,2}' --include="*.py" --include="*.js" --include="*.ts" --include="*.md" \
  --exclude-dir={node_modules,vendor,.git} . 2>/dev/null | head -10

# ANSI escape sequences
grep -rPn '\x1b\[|\x1b\]|\x1b\(' --exclude-dir={node_modules,vendor,.git} . 2>/dev/null | head -10

# Null bytes
grep -rPn '\x00' --exclude-dir={node_modules,vendor,.git,dist} . 2>/dev/null | head -5

# Nested command execution in markdown/config
grep -rn -E '\$\([^)]+\)|`[^`]+`' --include="*.md" --include="*.yaml" --include="*.json" \
  --exclude-dir={node_modules,vendor,.git} . 2>/dev/null | head -10
```

**Scoring:**
- 0 injection vectors → +15 points
- 1-2 vectors (likely false positives) → +10 points
- 3+ vectors → +5 points
- Confirmed injection in CLAUDE.md → 0 points

---

### Phase 4: Dependency audit

Run the appropriate package audit for the project:

```bash
# Node.js
[ -f "package-lock.json" ] && npm audit --json 2>/dev/null | jq '{total: .metadata.vulnerabilities.total, critical: .metadata.vulnerabilities.critical, high: .metadata.vulnerabilities.high}' 2>/dev/null

# Python
[ -f "requirements.txt" ] && pip-audit -r requirements.txt 2>/dev/null || [ -f "pyproject.toml" ] && pip-audit 2>/dev/null

# Rust
[ -f "Cargo.toml" ] && cargo audit 2>/dev/null

# Go
[ -f "go.mod" ] && govulncheck ./... 2>/dev/nu

CAUTIONgrade B · trust 89/100 Install with care. The audit found things worth knowing before you trust its output.

What the source does

Filesystem

none-observed

Network

none-observed

Shell

none-observed

Dependencies

pinned

Secrets in source

none-found

Findings (4)

whitepapers/recap-cards/en/_extensions

whitepapers/recap-cards/fr/_extensions

grep -rn '<!--' --include="*.md" . 2>/dev/null | grep -i 'ignore\|system\|admin\|instruction\|override\|forget'

You are a senior application security engineer. Perform a 6-phase security audit and produce a scored report with prioritized remediation plan.

Gates applied: no_behavioural_pass.