skillforge
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@skillforgesearch for a code review skill"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
SkillForge
SkillForge is a dependency-light MVP for creating, validating, discovering, installing, and packaging portable SKILL.md files. The universal format uses YAML frontmatter with required name and description fields followed by Markdown instructions.
Documentation: https://gintingism.github.io/skillforge/
Install
python -m pip install agent-skillforgeThe CLI is exposed as skillforge (or python -m skillforge.cli in an environment configured for module execution).
Related MCP server: Skills MCP Server
Usage
skillforge create code-review --description "Review code carefully" --directory .\skills\code-review
skillforge validate .\skills\code-review\SKILL.md
skillforge validate .\skills\code-review\SKILL.md --json
skillforge pack .\skills\code-review --output .\code-review.zip
skillforge search review --registry .\.skillforge
skillforge install code-review --registry .\.skillforge --project .The local registry stores skills as <registry>/<name>/SKILL.md. The explicit built-in registry ships seven skills and acts as fallback after configured local skills. The registry interface is intentionally abstract so a GitHub-backed implementation can be added without changing CLI consumers. Installation detects .github for Copilot, .claude for Claude Code, and .cursor for Cursor. Copilot receives .github/copilot-instructions.md; Claude Code receives .claude/skills/<name>/SKILL.md; Cursor receives .cursor/skills/<name>/SKILL.md.
validate --json emits machine-readable validation results without changing human-readable output or exit codes.
MCP server
SkillForge uses the official MCP Python SDK. Install project dependencies, then configure an MCP client to run:
{
"mcpServers": {
"skillforge": {
"command": "skillforge",
"args": ["serve-mcp"],
"env": {
"SKILLFORGE_REGISTRY_PATH": "C:\\path\\to\\.skillforge"
}
}
}
}The server exposes skill://{name} resources and read-only tools search_skills, get_skill, and validate_skill. Protocol messages use stdout; server logs use stderr. server.json contains the same launcher metadata.
Development
python -m pip install -e ".[dev]"
python -m pytest
python -m mypy src
mkdocs build --strictThis server cannot be deployed
Maintenance
Related MCP Connectors
Search and discover Agent Skills from the skills.sh registry. Powered by HAPI MCP server.
Discover and inspect public Runx skills before installing them. This read-only hosted MCP exposes search_skills and get_skill. The Apache-2.0 Rust runtime executes skills locally with scoped authority, approval gates and signed receipts. Docs: https://runx.ai/docs/mcp.
Search verified Claude Code plugins and skills; fetch portable SKILL.md sources. Read-only.
Search, fetch, lint, and install Agent Skills (SKILL.md) from the SkillMD registry.
Related MCP Servers
- AlicenseAqualityDmaintenanceExposes Cursor-style skill packs (SKILL.md trees) via MCP resources, tools, and prompts, enabling any MCP client to list, get, search skills, and retrieve skill context.34 npmISC
- AlicenseNot gradedqualityCmaintenanceEnables MCP clients to infer and resolve skills from text, normalize skill names, perform semantic taxonomy searches, and inspect lifecycle governance.MIT
- FlicenseCqualityCmaintenanceExposes Spec-Driven Development specs and reusable skills to MCP clients, enabling them to validate specs, list and fetch skills, and generate ready-to-use prompts for code generation.16-
- AlicenseNot gradedqualityBmaintenanceEnables AI clients to discover, search, retrieve, and read supporting files from a centrally published, versioned skill registry through four read-only MCP tools.1,505 npmMIT