AI-powered skill management system that enables AI agents to create, load, evolve, and manage reusable skills with market integration, smart dispatch, and version control.
Skillix (Skill + Mix) is a Model Context Protocol (MCP) based skill management system that provides AI coding agents with the ability to create, manage, load, and evolve reusable skills. It features intelligent task dispatch, marketplace integration, local-first architecture, and version control.
This skill integrates the Skillix MCP server into your AI agent environment, enabling:
Add Skillix to your AI agent's MCP configuration:
**For Cursor**: Edit `~/.cursor/mcp.json`
```json
{
"mcpServers": {
"skillix": {
"command": "npx",
"args": ["skillix-mcp"]
}
}
}
```
**For Claude Desktop**: Edit appropriate config file for your platform
**For other agents**: Follow their MCP server configuration guidelines
After MCP server is connected, initialize Skillix for your project:
```
Please initialize Skillix for this project using sx-config action=init
```
This creates:
Check that tools are available:
```
Please list available Skillix tools
```
You should see: `sx-dispatch`, `sx-skill`, `sx-market`, `sx-config`, `sx-feedback`, `sx-help`
Use `sx-dispatch` as your primary entry point for task analysis:
**Basic dispatch:**
```
sx-dispatch task="Convert PDF files to images"
```
**With context:**
```
sx-dispatch task="Process Excel data" context="Python project with pandas" hints=["excel", "data-processing"]
```
**Dispatch action types:**
**Example workflow:**
1. AI receives task from user
2. Call `sx-dispatch` with task description
3. Receive recommended action and skill name
4. Use `sx-skill read` to load skill instructions
5. Execute task following skill guidance
**List all skills:**
```
sx-skill action=list
```
**Read a skill:**
```
sx-skill action=read name=pdf-converter
```
**Create new skill:**
```
sx-skill action=create name=my-skill metadata={"name":"my-skill","description":"Process data files","version":"1.0.0","author":"AI","tags":["data","processing"]} body="# My Skill\n\n## Instructions\n\nDetailed skill instructions here..."
```
**Update skill metadata:**
```
sx-skill action=update name=my-skill metadata={"version":"1.1.0","tags":["data","processing","improved"]}
```
**Update skill content:**
```
sx-skill action=update name=my-skill body="# Updated Skill\n\nNew instructions..."
```
**Delete skill:**
```
sx-skill action=delete name=my-skill
```
**Search for skills:**
```
sx-market action=search query=pdf
```
**Install from marketplace:**
```
sx-market action=install name=pdf-converter scope=global
```
**Force overwrite existing:**
```
sx-market action=install name=pdf-converter force=true
```
**Uninstall skill:**
```
sx-market action=uninstall name=pdf-converter
```
**Sync market sources:**
```
sx-market action=sync
```
**Check source status:**
```
sx-market action=status
```
**View configuration:**
```
sx-config action=get scope=global
```
**Initialize project:**
```
sx-config action=init projectRoot=/path/to/project
```
**Refresh Cursor Rule:**
```
sx-config action=refresh projectRoot=/path/to/project
```
**Add skill source:**
```
sx-config action=sources sourceAction=add source={"name":"custom-source","url":"https://github.com/user/skills"}
```
**List sources:**
```
sx-config action=sources sourceAction=list
```
**Remove source:**
```
sx-config action=sources sourceAction=remove sourceName=custom-source
```
**Record success:**
```
sx-feedback action=record skillName=my-skill result=success task="Successfully processed 100 files"
```
**Record failure:**
```
sx-feedback action=record skillName=my-skill result=failure notes="Failed with error: timeout"
```
**List feedback:**
```
sx-feedback action=list skillName=my-skill days=7
```
**Analyze & get suggestions:**
```
sx-feedback action=analyze skillName=my-skill
```
**Clear feedback:**
```
sx-feedback action=clear skillName=my-skill
```
**Get help on specific tool:**
```
sx-help topic=dispatch
```
**Available help topics:**
Skills are Markdown files with YAML frontmatter:
```markdown
---
name: my-skill
description: What this skill does
version: 1.0.0
author: creator-name
tags: [tag1, tag2, tag3]
---
Clear description of what this skill accomplishes.
Step-by-step instructions for executing this skill.
Practical usage examples.
Common issues and solutions.
```
```
my-skill/
├── SKILL.md # Required: Main skill definition
├── scripts/ # Optional: Helper scripts
├── references/ # Optional: Documentation
├── assets/ # Optional: Resources
├── logs/ # Optional: Execution logs
└── .backup/ # Auto-generated: Version backups
```
1. **Skill lookup**: Project skills → Global skills → Remote market
2. **Configuration**: Project config → Global config → Defaults
3. **Same-name skills**: Project-level overrides global
```
~/.skillix/
├── config.json # Global configuration
├── skills/ # Global skill library
├── installed.json # Installation tracking
├── logs/ # System logs
├── cache/ # Cached repositories
└── data/ # Application data
```
```
project/
├── .skillix/
│ ├── config.json # Project-specific config
│ ├── skills/ # Project-specific skills
│ └── logs/ # Project logs
└── .cursor/
└── rules/
└── skillix.mdc # Auto-installed Cursor Rule
```
```
1. User provides task
2. AI calls: sx-dispatch task="<description>"
3. Dispatch recommends: USE_EXISTING with skill name
4. AI calls: sx-skill action=read name=<skill-name>
5. AI executes task following skill instructions
6. AI calls: sx-feedback action=record with results
```
```
1. User requests functionality not in local skills
2. AI calls: sx-dispatch task="<description>"
3. Dispatch recommends: INSTALL with skill name
4. AI calls: sx-market action=install name=<skill-name>
5. AI calls: sx-skill action=read name=<skill-name>
6. AI executes task
```
```
1. User requests new capability
2. AI calls: sx-dispatch task="<description>"
3. Dispatch recommends: CREATE_NEW
4. AI designs skill content
5. AI calls: sx-skill action=create with metadata and body
6. AI uses newly created skill
7. AI records feedback for future improvements
```
```
1. AI receives feedback on skill performance
2. AI calls: sx-feedback action=analyze skillName=<name>
3. Review suggestions and current skill
4. AI calls: sx-skill action=update with improvements
5. Old version automatically backed up to .backup/
```
1. Verify Node.js version: `node --version` (must be >= 18.0.0)
2. Check MCP config file location and JSON syntax
3. Restart AI agent after configuration changes
4. Clear npm cache if using npx: `npm cache clean --force`
```bash
npm cache clean --force
rm -rf ~/.npm/_npx/*
```
```bash
sudo chown -R $(whoami) ~/.npm
```
1. Check skill naming follows hyphen-case rules
2. Verify SKILL.md exists in skill directory
3. Validate YAML frontmatter syntax
4. Confirm project initialization: `sx-config action=get`
1. **Always use sx-dispatch first** for task analysis before manually selecting skills
2. **Record feedback** after skill execution to enable continuous improvement
3. **Use project-level skills** for project-specific workflows
4. **Keep global skills** for general, reusable capabilities
5. **Regular market sync** to discover new community skills
6. **Version skills properly** using semantic versioning (1.0.0, 1.1.0, 2.0.0)
7. **Tag skills appropriately** for better discoverability
8. **Document prerequisites** clearly in skill content
Leave a review
No reviews yet. Be the first to review this skill!
# Download SKILL.md from killerskills.ai/api/skills/skillix-mcp-skill-manager/raw