Agent skill
maintaining-claude-code
Create, validate, and improve Claude Code configuration — SKILL.md files, CLAUDE.md, rules, hooks, and settings.json. Use when creating a new skill, writing a SKILL.md, adding a hook, editing rules, auditing skill descriptions, checking config quality, debugging hook behavior, or deciding between skills vs rules vs CLAUDE.md. Also auto-loads when working in ~/.claude/ on skills, rules, hooks, or settings.
Install this agent skill to your Project
npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/other/other/maintaining-claude-code
SKILL.md
Maintaining Claude Code
Validate, organize, and improve Claude Code configurations.
Modes of Operation
Audit Mode
Use when: Checking config quality, validating skills work
Checklist:
- CLAUDE.md: Specific, structured, actionable
- Skills: Valid YAML, good descriptions (What + When formula)
- Commands: Clear purpose, not duplicating skills
- Hooks: Proper exit codes, reasonable timeouts
Organize Mode
Use when: .claude directory is messy, too many similar skills
Guidelines:
- Split CLAUDE.md into rules when >150 lines
- Consolidate similar skills (don't have 3 "code review" skills)
- Use subdirectories in rules/ for large projects
Advise Mode
Use when: Deciding what entity type to create
Decision tree:
- Needs to run automatically before/after actions? -> Hook
- Claude should auto-detect and use? -> Skill
- Needs isolated context for heavy work? -> Skill with
context: fork - Always-on behavioral guidance? -> CLAUDE.md
- Path-specific rules? -> .claude/rules/
Quick Reference
Entity Type Decision Matrix
| Need | Best Entity | Alternative |
|---|---|---|
| Global behavior guidelines | CLAUDE.md | Rules if >150 lines |
| Path-specific rules | .claude/rules/ | CLAUDE.md if universal |
| Auto-detected capabilities | Skills | Rules if always-on |
| Heavy isolated workflows | Skills (context: fork) | Regular skill |
| Pre/post action validation | Hooks | Nothing else does this |
| External API integration | MCP Servers | Bash calls if simple |
Skill Description Formula
<What it does>. Use when <trigger1>, <trigger2>, or <trigger3>.
Good: "Extract text and tables from PDF files. Use when working with PDFs, forms, or document extraction."
Bad: "Helps with documents"
YAML Validation
---on line 1 (required)name:max 64 charsdescription:max 1024 chars, must include triggers---before content
Common Anti-Patterns
- Vague descriptions: "Helps with stuff"
- Nested references: SKILL.md -> REF.md -> DETAILS.md
- Overloaded skills: Does 5 unrelated things
- Missing triggers: No "Use when..." clause
Validation Steps
- Check YAML syntax in all skills
- Verify descriptions include trigger phrases
- Ensure no duplicate capabilities across skills
- Confirm CLAUDE.md content won't quickly grow stale
- Check hooks have reasonable timeouts
Resources
See REFERENCE.md for detailed examples and troubleshooting.
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
agent-ops-spec
Manage specification documents in .agent/specs/. Use when user provides requirements, acceptance criteria, or feature descriptions that need to be tracked and validated against implementation.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-testing
Test strategy, execution, and coverage analysis. Use when designing tests, running test suites, or analyzing test results beyond baseline checks.
agent-ops-state
Maintain .agent state files. Use at session start, after meaningful steps, and before concluding: read/update constitution/memory/focus/issues/baseline consistently.
Didn't find tool you were looking for?