Skill Router
Find the right Claude Code skill from a large catalog when your request is vague or you are not sure which specialized skill to invoke.
Overview
skill-router is a journey-wide agent skill that maps natural-language tasks to the best Claude Code skill in your catalog—usable whenever a solo builder needs to pick the right capability before committing to a workflow.
Install
npx skills add https://github.com/charon-fan/agent-playbook --skill skill-routerWhat is this skill?
- Four-step routing process: intent analysis, skill matching, interactive clarification, and recommendation with reasoning
- Handles direct matches (e.g. PR review → code-reviewer) and ambiguous 'help with my project' requests via targeted quest
- Designed for symlink install into ~/.claude/skills alongside agent-playbook skills
- Semantic and keyword matching against the available skill catalog before activation
- Offers to activate the recommended skill after presenting the match
- 4-step routing process documented in How It Works
Adoption & trust: 710 installs on skills.sh; 58 GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
You have many agent skills installed but do not know which one matches your task, so you default to generic chat instead of procedural skills.
Who is it for?
Builders with a growing ~/.claude/skills tree who want intent-based discovery without memorizing every skill slug.
Skip if: Single-skill projects, fully specified invocations where you already know the exact skill name, or environments with no skill catalog to match against.
When should I use this skill?
When the user wants help picking or activating a skill from the catalog, or asks to 'use a skill' without naming one.
What do I get? / Deliverables
You receive a named skill recommendation with reasoning—and optional clarification questions—so you can activate the right playbook skill for the job.
- Recommended skill name with short justification
- Clarifying questions when intent is ambiguous
- Optional activation of the chosen skill
Recommended Skills
Journey fit
Useful at every journey phase - explore requirements and options before committing to a direction.
Where it fits
Ask how to explore competitors and get routed to a research-oriented skill instead of generic brainstorming.
Describe a UI change and receive frontend-design or a related component skill recommendation.
Say you need to review a pull request and activate code-reviewer with a clear match path.
Report a production bug and get steered toward debugging or monitoring skills in the catalog.
Ask for help scoping an MVP when multiple planning skills exist and clarification questions narrow the pick.
How it compares
Meta routing over your local skill catalog—not an MCP server registry and not a substitute for running the recommended skill’s full workflow.
Common Questions / FAQ
Who is skill-router for?
Claude Code users maintaining multiple skills from agent-playbook or similar repos who want automated skill selection from plain English requests.
When should I use skill-router?
Across Idea, Validate, Build, Ship, Launch, Grow, and Operate whenever you say 'use a skill,' need PR review help, or any task where the correct specialized skill is unclear—router runs before the downstream skill.
Is skill-router safe to install?
Check the Security Audits panel on this Prism page; installation involves local symlinks under ~/.claude/skills—review paths and repo trust before linking.
SKILL.md
READMESKILL.md - Skill Router
# Skill Router An intelligent routing layer for Claude Code skills that helps users find the right skill for their task. ## Overview The Skill Router analyzes user requests and intelligently recommends the most appropriate skill from the available catalog. It uses intent classification, semantic matching, and interactive clarification to ensure users get the help they need. ## Installation ```bash mkdir -p ~/.claude/skills ln -s ~/Documents/code/GitHub/agent-playbook/skills/skill-router ~/.claude/skills/skill-router ``` ## How It Works ``` User Request → Intent Analysis → Skill Matching → Recommendation ↓ Interactive Clarification (if needed) ``` ### Routing Process 1. **Intent Analysis**: Parse user request to identify task type, context, and complexity 2. **Skill Matching**: Match intent against available skills using keywords and semantics 3. **Interactive Clarification**: Ask targeted questions if request is ambiguous 4. **Recommendation**: Present the best matching skill with reasoning ## Usage Examples ### Direct Skill Match ``` You: "I need to review this pull request" Skill Router: Recommended Skill: code-reviewer This skill is designed for reviewing PRs and code changes for quality, security, and best practices. Would you like me to activate this skill? ``` ### Ambiguous Request ``` You: "Use a skill to help with my project" Skill Router: I'd like to understand your needs better: - What type of task are you working on? - Are you designing, coding, testing, or documenting? - What's the primary goal? [Based on your answers, I'll recommend the right skill] ``` ### Multi-Skill Workflow ``` You: "I'm building a new API from scratch" Skill Router: For a complete API project, I recommend using multiple skills: 1. api-designer - Design your API structure and endpoints 2. api-documenter - Create OpenAPI/Swagger documentation 3. test-automator - Set up automated API testing 4. code-reviewer - Review your implementation Shall we start with api-designer? ``` ## Supported Skills | Category | Skills | |----------|--------| | **Core Development** | commit-helper, code-reviewer, debugger, refactoring-specialist | | **Documentation & Testing** | documentation-engineer, api-documenter, test-automator, qa-expert | | **Architecture & DevOps** | api-designer, security-auditor, performance-engineer, deployment-engineer | | **Planning & Analysis** | architecting-solutions, planning-with-files, long-task-coordinator, self-improving-agent | ## Features - **Intent Classification**: Understands what the user wants to accomplish - **Semantic Matching**: Finds relevant skills even with different wording - **Interactive Guidance**: Asks clarifying questions when needed - **Multi-Skill Orchestrations**: Suggests skill combinations for complex tasks - **Confidence Indicators**: Shows how confident the recommendation is ## Best Practices ### For Users 1. **Be specific**: "I need to write tests for my API" vs "Help with tests" 2. **Provide context**: Mention your project type and stage 3. **Answer clarification questions**: Helps the router find the best match 4. **Request alternatives**: Ask "What are my options?" for multiple choices ### For Maintainers 1. **Update skill catalog**: Add new skills to the catalog table 2. **Refine routing logic**: Add new patterns and keywords 3. **Track edge cases**: Note requests that don't fit existing skills 4. **Improve questions**: Refine clarification templates based on usage ## Routing Patterns | User Says | Routes To | |-----------|-----------| | "review my code" | code-reviewer | | "fix this bug" | debugger | | "write tests" | test-automator | | "create documentation" | documentation-engineer | | "design an API" | api-designer | | "deploy to production" | deployment-engineer | | "improve performance" | performance-engineer | | "check security" | security-auditor | | "refactor code" | refactoring-specialist | | "res