Skills-ContextManager
Don’t pollute your AI agent’s context with 1,000 skills. Use Skills-ContextManager, a self-hosted web UI for managing AI skills and workflows by providing skills to an AI agent via MCP only when needed. Simply add skills to your library and enable or disable them with a toggle. Choose whether a skill is always loaded into context or dynamically activated when the AI agent determines it’s needed.
Skill-ContextManager & MCP Server
A self-hosted web application for managing AI skills, workflows, and contexts with full MCP (Model Context Protocol) integration. Organize, manage, and dynamically load specialized knowledge bases into any AI Agent just by toggling your Skills On/Off in simple local hosted WEB UI.

Features
Web Application
- Skill & Workflow Management: Create, edit, delete, and organize skills and workflows
- Multi-Hub Support: Create isolated workspaces with separate skills/contexts
- Context Cells: Organize skills/workflows into logical groups with enable/disable toggles
- Dual Mode: Choose Always-loaded (auto-loaded) or Dynamic (on-demand) skills by on/off toggle
- File Editor: Built-in editor with file tree, create/delete files & folders
- Import Sources:
- 📂 Folder: Upload from local filesystem
- 🐙 GitHub: Clone from GitHub repositories
- ⚡ Skills.sh: Import from Skills.sh registry
- AI Description Generation: Auto-generate skill descriptions using AI APIs
- Secure API Key Storage: API keys stored with file mode 0600 (owner-only)
- Settings Profiles: Save and load configuration presets
- Search & Filter: Smart search, sort, favorites, active/inactive filters
- Token Counting: Track context size in tokens
- Drag & Drop: Move skills/workflows to contexts
MCP Server
- 3 Essential Tools for AI Agents:
get_default_skills()- Load skills with mode"always_loaded"list_available_skills()- List all enabled skills with modesload_full_skill_context(name)- Load dynamic skill on-demand
- Hub-Aware: Works with active hub from multiple workspaces
- Mode Support:
"always_loaded"vs"dynamic"skill loading - Type Support: Both Skills and Workflows
Quick Start
# Clone the repository
git clone https://github.com/One-Man-Company/Skills-ContextManager.git
cd Skills-ContextManager
# Make setup script executable and run it
chmod +x setup.sh
./setup.sh
# Start web application
cd Skill-ContextManager && npm start
Access: http://localhost:3000
MCP configuration saved to mcp_settings.json - add this to your AI Agent's MCP settings.
⚠️ Important: Copy the content of
MCP_instructions.mdto your AI Agent's System prompt or rules configuration.
MCP Configuration
The setup.sh script generates mcp_settings.json with correct paths. Add it to your AI Agent's MCP settings file:
{
"Skills-ContextManager": {
"command": "/absolute/path/Skills-MCP/.venv/bin/python3",
"args": ["-u", "/absolute/path/Skills-MCP/mcp_server.py"],
"env": {}
}
}
MCP Tools
AI Agents need only these 3 tools. Usage details are in MCP_instructions.md (system prompt):
| Tool | Description |
|---|---|
get_default_skills() | Load all skills with mode "always_loaded" into context |
list_available_skills() | List all enabled skills with descriptions and modes |
load_full_skill_context(name) | Load a specific skill on-demand (for mode "dynamic") |
mcpservers.org Submission
| Field | Value |
|---|---|
| Name | Skills-ContextManager |
| Description | Web UI for managing AI skills with MCP integration |
| Repository | https://github.com/One-Man-Company/Skills-ContextManager |
| Command | [path]/Skills-MCP/.venv/bin/python3 |
| Args | -u [path]/Skills-MCP/mcp_server.py |
Included Skills & Workflows
The repository includes starter content in MySkillsHUB/:
Skills (39): architecture, behavioral-modes, clean-code, code-review-checklist, documentation-templates, frontend-design, game-development, i18n-localization, mobile-design, nextjs-react-expert, nodejs-best-practices, parallel-agents, performance-profiling, python-patterns, seo-fundamentals, server-management, skill-creator, systematic-debugging, ui-ux-pro-max, vulnerability-scanner, webapp-testing, api-patterns, and more.
Workflows (3):
- best-standard-autonomous-coding
- elite-cognitive-operator-system
- plan-mode
All content is automatically copied to your library on first setup.
License
MIT License
Created: February 2026
相關伺服器
Alpha Vantage MCP Server
贊助Access financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
Remote MCP Server (Authless)
An example of a remote MCP server without authentication, deployable on Cloudflare Workers.
Hackle
Query A/B test data using the Hackle API.
LSP MCP Server
Integrates with Language Server Protocol (LSP) to provide features like code completion, diagnostics, and hover information.
DeepView MCP
Enables IDEs like Cursor and Windsurf to analyze large codebases using Gemini's 1M context window.
velixar-mcp-server
Persistant AI Memory
TrueNAS Middleware MCP Server
Accesses optimized documentation from the TrueNAS middleware repository to understand its codebase and APIs.
FastAPI-MCP
A zero-configuration tool to automatically expose FastAPI endpoints as MCP tools.
MCP Create Server
A service for dynamically creating, running, and managing Model Context Protocol (MCP) servers.
Qwen-Agent
A framework for developing LLM applications with capabilities like tool usage, planning, and memory, based on the Qwen model.
MCP Shell
Execute secure shell commands from AI assistants and other MCP clients, with configurable security settings.