Add WinDAGs to the tool
you already use.
Your agent gets a library of expert skills and a way to plan work before it runs. Start simple; the technical install paths are below.

What you'll have when you're done
- Your AI agent can search and use specialist skills when the work needs real domain knowledge.
- In Claude Code: a
/next-moveslash command that predicts your next 3-wave action plan from your git state and recent files. - No API keys for the read-only install path. Search and skill lookup run locally; execution uses whatever provider you configure in your client.
Prerequisites
- One of: Claude Code, Claude Desktop, Cursor, Codex CLI, Gemini CLI, or any MCP-compatible client.
- Node.js 18 or later (only for the MCP path; Claude Code path doesn't need it).
- About 30 seconds.
Pick your path
Two ways to install. Pick whichever matches your client. They give you slightly different surfaces (slash commands vs MCP tools) and you can install both.
You get slash commands (/next-move), the meta-DAG planning agents, and the full skill catalog landing in ~/.claude/. Everything appears in your tool picker.
Universal install for any client that speaks Model Context Protocol. Your agent gets three tools: skill_search, skill_graft, skill_reference.
Plugin marketplace (Claude Code)
The fastest install. Adds the marketplace, installs the plugin, and you're done. Slash commands and skills appear automatically in Claude Code.
$ claude plugin marketplace add curiositech/windags-skills && claude plugin install windags-skills
~/.claude/skills/ and ~/.claude/agents/./next-move. If it autocompletes, you're set.MCP server (any client)
The MCP server exposes windags_skill_search, windags_skill_graft, and windags_skill_reference over stdio. Zero API keys for the read-only tools — BM25 runs locally.
Claude Code
$ claude mcp add windags -- npx -y @workgroup-ai/mcp-server
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"windags": {
"command": "npx",
"args": ["-y", "@workgroup-ai/mcp-server"]
}
}
}Cursor
Edit ~/.cursor/mcp.json:
{
"mcpServers": {
"windags": {
"command": "npx",
"args": ["-y", "@workgroup-ai/mcp-server"]
}
}
}Codex / Gemini CLI / generic stdio MCP
Any client that supports stdio MCP works. Run npx @workgroup-ai/mcp-server and connect over stdio.
Verify
- Open a session in your client and ask: "search WinDAGs for skills about Postgres connection pooling". The agent should call
windags_skill_searchand return ranked matches. - Then: "graft the top match into context". The full skill body should appear and the agent's next answer should be visibly more specific.
- For Claude Code only — type
/next-moveto invoke the prediction pipeline.
Privacy
Telemetry is local-only. Every skill graft and reference load gets recorded to ~/.windags/skill-state.db (SQLite) on your machine. The data feeds local re-ranking and gap detection. There is no --share flag yet; when one ships it will be opt-in, hashed before upload, and documented in public before the flag goes live.
Troubleshooting
Run one command. Keep what helps. Uninstall in one second if it doesn't.
No account. No tracking pixel. No phone-home. Telemetry stays on your machine. Uninstall is claude plugin uninstall windags-skills.
$ claude plugin marketplace add curiositech/windags-skills && claude plugin install windags-skills