MCP Tools Reference

Skill Seekers provides 26 MCP (Model Context Protocol) tools for Claude Code Desktop and other AI agents. These tools enable natural language commands for skill creation, codebase analysis, and knowledge management.

Overview

MCP tools allow AI agents to:

  • Create skills from documentation, code, and PDFs
  • Analyze codebases with C3.x pattern detection
  • Enhance knowledge with AI-powered improvements
  • Manage configurations for different platforms

All tools work across 5 AI agent platforms: Claude Code, Cursor, Windsurf, VS Code, and IntelliJ.

Tool Categories

🎯 Skill Creation Tools (8)

Tools for creating skills from various sources:

ToolDescriptionExample Command
scrape_docsScrape documentation websites”Scrape React docs”
scrape_githubAnalyze GitHub repositories”Analyze facebook/react”
scrape_pdfExtract content from PDFs”Extract my-api.pdf”
unified_scrapeMulti-source skill creation”Create skill from docs + GitHub”
create_skillCreate skill from local files”Create skill from ./my-project”
clone_skillDuplicate existing skill”Clone my-react skill”
merge_skillsCombine multiple skills”Merge react and redux skills”
split_skillSplit large skills”Split skill into smaller chunks”

🔍 Analysis Tools (6)

C3.x codebase analysis capabilities:

ToolDescriptionExample Command
analyze_codebaseFull codebase analysis”Analyze this codebase”
detect_patternsFind design patterns”Detect patterns in src/“
extract_testsPull test examples”Extract test examples”
generate_guidesCreate how-to guides”Generate setup guide”
analyze_configConfig pattern extraction”Analyze config files”
generate_architectureCreate ARCHITECTURE.md”Generate architecture doc”

✨ Enhancement Tools (4)

AI-powered skill improvements:

ToolDescriptionExample Command
enhance_skillAI-enhance a skill”Enhance my skill”
improve_descriptionsBetter skill descriptions”Improve descriptions”
add_examplesGenerate code examples”Add Python examples”
validate_skillCheck skill quality”Validate this skill”

📦 Packaging Tools (4)

Platform-specific packaging:

ToolDescriptionExample Command
package_for_claudePackage for Claude AI”Package for Claude”
package_for_langchainPackage for LangChain”Package for LangChain”
package_for_llamaindexPackage for LlamaIndex”Package for LlamaIndex”
package_for_codingPackage for AI IDEs”Package for Cursor”

⚙️ Configuration Tools (4)

Management and utilities:

ToolDescriptionExample Command
list_configsShow available configs”List all configs”
validate_configCheck config syntax”Validate my config”
estimate_jobEstimate time/cost”Estimate react docs scrape”
resume_jobResume interrupted job”Resume last job”

Detailed Tool Reference

scrape_docs

Scrape documentation websites into structured skills.

Parameters:

{
  "url": "https://react.dev",
  "max_pages": 100,
  "selectors": { "content": "article" },
  "output_dir": "./output/react"
}

Example Commands:

  • “Scrape React documentation”
  • “Scrape https://vuejs.org with 50 pages max”
  • “Scrape Django docs using custom selectors”

Returns: Path to created skill directory

scrape_github

Analyze GitHub repositories for code patterns and structure.

Parameters:

{
  "repo": "facebook/react",
  "include_issues": false,
  "include_tests": true,
  "output_dir": "./output/react-github"
}

Example Commands:

  • “Analyze facebook/react”
  • “Scrape GitHub repo with tests included”
  • “Analyze python/cpython with issues”

Returns: Path to created skill directory

analyze_codebase

Full C3.x analysis suite for local codebases.

Parameters:

{
  "directory": "./my-project",
  "comprehensive": true,
  "output_format": "claude"
}

Example Commands:

  • “Analyze this codebase comprehensively”
  • “Analyze src/ for patterns”
  • “Run C3.x suite on my project”

Returns: Analysis results with patterns, tests, and guides

enhance_skill

AI-enhance skill content for better quality.

Parameters:

{
  "skill_path": "./output/react",
  "method": "local",
  "platform": "claude"
}

Example Commands:

  • “Enhance my react skill”
  • “Use API enhancement for this skill”
  • “Enhance with Gemini instead of Claude”

Returns: Enhanced skill path

package_for_claude

Package skill for Claude AI platform.

Parameters:

{
  "skill_path": "./output/react",
  "include_router": true
}

Example Commands:

  • “Package for Claude”
  • “Create Claude skill with router”
  • “Package my skill for Claude AI”

Returns: Packaged skill ready for upload

estimate_job

Estimate time and cost before running.

Parameters:

{
  "config": "configs/react.json",
  "include_enhancement": true
}

Example Commands:

  • “Estimate React docs scrape”
  • “How long will this take?”
  • “Estimate cost for this job”

Returns: Time estimate and cost breakdown

Platform-Specific Features

Claude Code

  • Transport: stdio (native)
  • Best For: Interactive skill creation
  • Features: Full 26-tool support

Cursor

  • Transport: HTTP
  • Best For: Codebase analysis
  • Features: 18 core tools

Windsurf

  • Transport: HTTP
  • Best For: Web development skills
  • Features: 18 core tools

VS Code

  • Transport: stdio
  • Best For: Development workflows
  • Features: 20 tools

IntelliJ

  • Transport: HTTP
  • Best For: Java/Kotlin projects
  • Features: 18 core tools

Common Workflows

Workflow 1: Quick Skill Creation

1. scrape_docs → Create skill from docs
2. enhance_skill → AI improvement
3. package_for_claude → Platform packaging
4. validate_skill → Quality check

Workflow 2: Codebase Analysis

1. analyze_codebase → Full analysis
2. detect_patterns → Find patterns
3. extract_tests → Get examples
4. generate_guides → Create tutorials
5. package_for_coding → IDE integration

Workflow 3: Multi-Source Skill

1. unified_scrape → Combine sources
2. merge_skills → Merge outputs
3. enhance_skill → AI enhancement
4. package_for_langchain → RAG pipeline

Troubleshooting

Tool Not Found

# Re-run MCP setup
./setup_mcp.sh

Permission Errors

# Check MCP settings file
ls -la ~/.claude/mcp_settings.json

Timeout Issues

  • Increase timeout in MCP settings
  • Use --timeout flag for long operations

See Also