health（技能安全扫描）

Claude Code Configuration Health Audit

Audit the current project's Claude Code setup with the six-layer framework: CLAUDE.md → rules → skills → hooks → subagents → verifiers

The goal is to find violations and identify the misaligned layer, calibrated to project complexity.

Output language: Check in order: (1) CLAUDE.md ## Communication rule (global takes precedence over local); (2) language of the user's recent conversation messages; (3) default English. Apply the detected language to all output including progress lines, the report, and the stop-condition question.

IMPORTANT: Before the first tool call, output a progress block in the output language:

Step 1/3: Collecting configuration data
  · CLAUDE.md (global + local) · rules/ · settings.local.json · hooks
  · MCP servers · skills inventory + security scan
  · conversation history (up to 3 recent sessions)

Step 0: Assess project tier

Pick tier:

Apply only the detected tier's requirements.

Step 1: Collect all data (single bash block)

Run one block to collect data.

P=$(pwd)
SETTINGS="$P/.claude/settings.local.json"

echo "=== TIER METRICS ==="
echo "project_files: $(git -C "$P" ls-files 2>/dev/null | wc -l || find "$P" -type f -not -path "*/.git/*" -not -path "*/node_modules/*" -not -path "*/dist/*" -not -path "*/build/*" | wc -l)"
echo "contributors: $(git -C "$P" log -n 500 --format='%ae' 2>/dev/null | sort -u | wc -l)"
echo "ci_workflows:  $(ls "$P/.github/workflows/"*.yml "$P/.github/workflows/"*.yaml 2>/dev/null | wc -l)"
echo "skills:        $(find "$P/.claude/skills" -name "SKILL.md" 2>/dev/null | grep -v '/health/SKILL.md' | wc -l)"
echo "claude_md_lines: $(wc -l < "$P/CLAUDE.md" 2>/dev/null)"

echo "=== CLAUDE.md (global) ===" ; cat ~/.claude/CLAUDE.md 2>/dev/null || echo "(none)"
echo "=== CLAUDE.md (local) ===" ; cat "$P/CLAUDE.md" 2>/dev/null || echo "(none)"
echo "=== settings.local.json ===" ; cat "$SETTINGS" 2>/dev/null || echo "(none)"
echo "=== rules/ ===" ; find "$P/.claude/rules" -name "*.md" 2>/dev/null | while IFS= read -r f; do echo "--- $f ---"; cat "$f"; done
echo "=== skill descriptions ===" ; { [ -d "$P/.claude/skills" ] && grep -r "^description:" "$P/.claude/skills" 2>/dev/null; grep -r "^description:" ~/.claude/skills 2>/dev/null; } | sort -u
echo "=== STARTUP CONTEXT ESTIMATE ==="
echo "global_claude_words: $(wc -w < ~/.claude/CLAUDE.md 2>/dev/null | tr -d ' ' || echo 0)"
echo "local_claude_words: $(wc -w < "$P/CLAUDE.md" 2>/dev/null | tr -d ' ' || echo 0)"
echo "rules_words: $(find "$P/.claude/rules" -name "*.md" 2>/dev/null | while IFS= read -r f; do cat "$f"; done | wc -w | tr -d ' ')"
echo "skill_desc_words: $({ [ -d "$P/.claude/skills" ] && grep -r "^description:" "$P/.claude/skills" 2>/dev/null; grep -r "^description:" ~/.claude/skills 2>/dev/null; } | wc -w | tr -d ' ')"
python3 -c "
import json, sys
try:
    d = json.load(open('$SETTINGS'))
except Exception as e:
    msg = '(unavailable: settings.local.json missing or malformed)'
    print('=== hooks ==='); print(msg)
    print('=== MCP ==='); print(msg)
    print('=== MCP FILESYSTEM ==='); print(msg)
    print('=== allowedTools count ==='); print(msg)
    sys.exit(0)

print('=== hooks ===')
print(json.dumps(d.get('hooks', {}), indent=2))

print('=== MCP ===')
s = d.get('mcpServers', d.get('enabledMcpjsonServers', {}))
names = list(s.keys()) if isinstance(s, dict) else list(s)
n = len(names)
print(f'servers({n}):', ', '.join(names))
est = n * 25 * 200
print(f'est_tokens: ~{est} ({round(est/2000)}% of 200K)')

print('=== MCP FILESYSTEM ===')
if isinstance(s, list):
    print('filesystem_present: (array format -- check .mcp.json)')
    print('allowedDirectories: (not detectable)')
else:
    fs = s.get('filesystem') if isinstance(s, dict) else None; a = []
    if isinstance(fs, dict):
        a = fs.get('allowedDirectories') or (fs.get('config', {}).get('allowedDirectories') if isinstance(fs.get('config'), dict) else [])
        if not a and isinstance(fs.get('args'), list):
            args = fs['args']
            for i, v in enumerate(args):
                if v in ('--allowed-directories', '--allowedDirectories') and i+1 < len(args): a = [args[i+1]]; break
            if not a: a = [v for v in args if v.startswith('/') or (v.startswith('~') and len(v) > 1)]
    print('filesystem_present:', 'yes' if fs else 'no')
    print('allowedDirectories:', a or '(missing or not detected)')

print('=== allowedTools count ===')
print(len(d.get('permissions', {}).get('allow', [])))
" 2>/dev/null || echo "(unavailable)"
echo "=== NESTED CLAUDE.md ===" ; find "$P" -maxdepth 4 -name "CLAUDE.md" -not -path "$P/CLAUDE.md" -not -path "*/.git/*" -not -path "*/node_modules/*" 2>/dev/null || echo "(none)"
echo "=== GITIGNORE ==="
_GITIGNORE_HIT=$(git -C "$P" check-ignore -v .claude/settings.local.json 2>/dev/null || true)
if [ -n "$_GITIGNORE_HIT" ]; then
  _GITIGNORE_SOURCE=${_GITIGNORE_HIT%%:*}
  case "$_GITIGNORE_SOURCE" in
    .gitignore|.claude/.gitignore)
      echo "settings.local.json: gitignored"
      ;;
    *)
      echo "settings.local.json: ignored only by non-project rule ($_GITIGNORE_SOURCE) -- add a repo-local ignore rule"
      ;;
  esac
else
  echo "settings.local.json: NOT gitignored -- risk of committing tokens/credentials"
fi
echo "=== HANDOFF.md ===" ; cat "$P/HANDOFF.md" 2>/dev/null || echo "(none)"
echo "=== MEMORY.md ===" ; cat "$HOME/.claude/projects/-$(pwd | sed 's|[/_]|-|g; s|^-||')/memory/MEMORY.md" 2>/dev/null | head -50 || echo "(none)"

echo "=== CONVERSATION FILES ==="
PROJECT_PATH=$(pwd | sed 's|[/_]|-|g; s|^-||')
CONVO_DIR=~/.claude/projects/-${PROJECT_PATH}
ls -lhS "$CONVO_DIR"/*.jsonl 2>/dev/null | head -10

echo "=== CONVERSATION EXTRACT (up to 3 most recent, confidence improves with more files) ==="
# Skip the active session, it may still be incomplete.
_PREV_FILES=$(ls -t "$CONVO_DIR"/*.jsonl 2>/dev/null | tail -n +2 | head -3)
if [ -n "$_PREV_FILES" ]; then
  echo "$_PREV_FILES" | while IFS= read -r F; do
    [ -f "$F" ] || continue
    echo "--- file: $F ---"
    head -c 2M "$F" | jq -r '
      if .type == "user" then "USER: " + ((.message.content // "") | if type == "array" then map(select(.type == "text") | .text) | join(" ") else . end)
      elif .type == "assistant" then
        "ASSISTANT: " + ((.message.content // []) | map(select(.type == "text") | .text) | join("\n"))
      else empty
      end
    ' 2>/dev/null | grep -v "^ASSISTANT: $" | head -300 || echo "(unavailable: jq not installed or parse error)"
  done
else
  echo "(no conversation files)"
fi

echo "=== MCP ACCESS DENIALS ==="
ls -t "$CONVO_DIR"/*.jsonl 2>/dev/null | head -5 | while IFS= read -r F; do
  head -c 1M "$F" | grep -Em 2 'Access denied - path outside allowed directories|tool-results/.+ not in ' 2>/dev/null
done | head -20

# --- Skill scan ---
# Exclude self by frontmatter name, stable across install paths.
SELF_SKILL=$( (grep -rl '^name: health$' "$P/.claude/skills" "$HOME/.claude/skills" 2>/dev/null || true) | grep 'SKILL.md' | head -1)
[ -z "$SELF_SKILL" ] && SELF_SKILL="health/SKILL.md"

echo "=== SKILL INVENTORY ==="
for DIR in "$P/.claude/skills" "$HOME/.claude/skills"; do
  [ -d "$DIR" ] || continue
  find -L "$DIR" -name "SKILL.md" 2>/dev/null | grep -v "$SELF_SKILL" | while IFS= read -r f; do
    WORDS=$(wc -w < "$f" | tr -d ' ')
    IS_LINK="no"; LINK_TARGET=""
    SKILL_DIR=$(dirname "$f")
    if [ -L "$SKILL_DIR" ]; then
      IS_LINK="yes"; LINK_TARGET=$(readlink -f "$SKILL_DIR")
    fi
    echo "path=$f words=$WORDS symlink=$IS_LINK target=$LINK_TARGET"
  done
done

echo "=== SKILL FRONTMATTER ==="
for DIR in "$P/.claude/skills" "$HOME/.claude/skills"; do
  [ -d "$DIR" ] || continue
  find -L "$DIR" -name "SKILL.md" 2>/dev/null | grep -v "$SELF_SKILL" | while IFS= read -r f; do
    if head -1 "$f" | grep -q '^---'; then
      echo "frontmatter=yes path=$f"
      sed -n '2,/^---$/p' "$f" | head -10
    else
      echo "frontmatter=MISSING path=$f"
    fi
  done
done

echo "=== SKILL SYMLINK PROVENANCE ==="
for DIR in "$P/.claude/skills" "$HOME/.claude/skills"; do
  [ -d "$DIR" ] || continue
  find "$DIR" -maxdepth 1 -type l 2>/dev/null | while IFS= read -r link; do
    TARGET=$(readlink -f "$link")
    echo "link=$(basename "$link") target=$TARGET"
    if [ -d "$TARGET/.git" ]; then
      REMOTE=$(git -C "$TARGET" remote get-url origin 2>/dev/null || echo "unknown")
      COMMIT=$(git -C "$TARGET" rev-parse --short HEAD 2>/dev/null || echo "unknown")
      echo "  git_remote=$REMOTE commit=$COMMIT"
    fi
  done
done

echo "=== SKILL FULL CONTENT (sample: up to 5 skills, 80 lines each) ==="
{ for DIR in "$P/.claude/skills" "$HOME/.claude/skills"; do
    [ -d "$DIR" ] || continue
    find -L "$DIR" -name "SKILL.md" 2>/dev/null | grep -v "$SELF_SKILL"
  done
} | head -5 | while IFS= read -r f; do
  echo "--- FULL: $f ---"
  head -80 "$f"
done

Gotchas

Before interpreting Step 1 output, check these known failure modes.

Data collection silent failures

• jq not installed: conversation extraction prints (unavailable: jq not installed or parse error). BEHAVIOR section will be empty -- treat as [INSUFFICIENT DATA], not a finding.
• python3 not on PATH: all MCP/hooks/allowedTools sections print (unavailable). Do not flag those areas when the data source itself failed.
• settings.local.json absent: hooks, MCP, and allowedTools all show (unavailable). Normal for projects using global settings only -- not a misconfiguration.

MEMORY.md path construction

• Path built with sed 's|[/_]|-|g' on pwd. Unusual characters produce the wrong project key. If MEMORY.md shows (none) but the user mentions prior sessions, verify the path manually before flagging as [!].

Conversation extract scope

• Only the 3 most recent .jsonl files are sampled, skipping the active session. Findings from fewer than 3 files carry low signal -- always tag [LOW CONFIDENCE].

MCP token estimate

• Assumes ~25 tools/server and ~200 tokens/tool. Servers with many or few tools cause large over/under-estimates. Treat as directional, not precise.

Tier misclassification edge cases

• The bash block excludes node_modules/, dist/, and build/, but not all generators. Monorepos with .next/, __pycache__/, or .turbo/ output can inflate the file count and trigger COMPLEX tier falsely. Recheck manually if the tier feels wrong.

Step 2: Analyze with tier-adjusted depth

After Step 1 completes, output a summary line, then the step indicator:

Tier: {SIMPLE/STANDARD/COMPLEX} -- {file_count} files · {contributor_count} contributors · CI: {present/absent}
Step 2/3: {SIMPLE: "Analyzing locally" | STANDARD/COMPLEX: "Launching parallel analysis agents"}

SIMPLE: output "Analyzing locally" above. Do not launch subagents. Analyze from Step 1, prioritize core config checks, skip conversation-heavy cross-validation unless evidence is obvious.

STANDARD/COMPLEX: output "Launching parallel analysis agents" above, then list coverage:

  · Agent 1: CLAUDE.md, rules, skills, MCP context + security scan
  · Agent 2: hooks, allowedTools, behavior patterns, three-layer defense

Launch two subagents in parallel. Paste all data inline -- do not pass file paths.

Agent 1 -- Context + Security Audit (no conversation needed)

Read agents/agent1-context.md from this skill's directory. It specifies which Step 1 sections to paste and the full audit checklist.

Agent 2 -- Control + Behavior Audit (uses conversation evidence)

Read agents/agent2-control.md from this skill's directory. It specifies which Step 1 sections to paste and the full audit checklist.

Step 3: Synthesize and present

Before writing the report, output a progress line in the output language:

Step 3/3: Synthesizing report

Aggregate the local analysis and any agent outputs into one report:

Health Report: {project} ({tier} tier, {file_count} files)

✓ Passing

Render a compact table of checks that passed. Include only checks relevant to the detected tier. Limit to 5 rows. Omit rows for checks that have findings.

☻ Critical -- fix now

Rules violated, missing verification definitions, dangerous allowedTools, MCP overhead >12.5%, required-path Access denied, active cache-breakers, and security findings.

◎ Structural -- fix soon

CLAUDE.md content that belongs elsewhere, missing hooks, oversized skill descriptions, single-layer critical rules, model switching, verifier gaps, subagent permission gaps, and skill structural issues.

○ Incremental -- nice to have

New patterns to add, outdated items to remove, global vs local placement, context hygiene, HANDOFF.md adoption, skill invoke tuning, and provenance issues.

If all three issue sections are empty, output one short line in the output language like: ✓ All relevant checks passed. Nothing to fix.

Non-goals

• Never auto-apply fixes without confirmation.
• Never apply complex-tier checks to simple projects.
• Flag issues, do not replace architectural judgment.

Stop condition: After the report, ask in the output language:

"Should I draft the changes? I can handle each layer separately: global CLAUDE.md / local CLAUDE.md / hooks / skills."

Do not make any edits without explicit confirmation.