Wavedash AI

AI integration assets for Wavedash.

The canonical skill source lives in the Wavedash docs repo under skills/ and is served from https://docs.wavedash.com/.well-known/skills/index.json. This repo syncs that source into skills/ so agent platforms can install or index the published assets from GitHub.

The mcp/ package contains a read-only, unauthenticated remote MCP server for public Wavedash docs. It does not access user accounts, local files, builds, or private Wavedash data.

Provider plugins include Wavedash skills and, where supported, bundled config for the hosted remote MCP server.

Layout

• skills/ - synced canonical skills for platforms that consume skills directly.

• providers/claude/plugin/ - Claude Code plugin package.

• providers/cursor/plugin/ - Cursor plugin package.

• providers/openai/plugin/ - Codex plugin package.

• mcp/ - read-only remote MCP server for public Wavedash docs.

• .claude-plugin/marketplace.json - Claude Code marketplace manifest.

• .cursor-plugin/marketplace.json - Cursor marketplace manifest.

• .agents/plugins/marketplace.json - Codex marketplace manifest.

• gemini-extension.json - Gemini CLI extension manifest using the root skills/ directory.

Related MCP server: Qdrant Docs MCP Server

Sync skills

Pull from the public docs endpoint:

npm run sync:skills

Pull from a local docs checkout:

WAVEDASH_SKILLS_SOURCE_DIR=/path/to/wvdsh/docs/skills npm run sync:skills

GitHub Actions runs the same sync daily and commits changes when docs updates change the generated skill files.

MCP server

The Wavedash remote MCP server is live at:

https://mcp.wavedash.com/mcp

It is read-only and unauthenticated. It exposes public Wavedash docs, SDK guidance, engine setup help, and publishing checklists for agents.

Remote MCP clients that support Streamable HTTP can connect directly to the endpoint above.

Run the docs-only MCP server locally:

cd mcp
npm install
npm start

The endpoint is http://localhost:3000/mcp.

Validate the Cloudflare Worker bundle without publishing:

npm run mcp:deploy:dry-run

Pushes to main automatically deploy the MCP Worker when MCP source, dependencies, Worker configuration, root package scripts, or the deployment workflow change. Documentation, skills, and submission files do not trigger a deployment. The Deploy MCP to Production GitHub Actions workflow also supports manual runs and loads deployment credentials through Doppler.

Deploy the MCP Worker directly from an environment that provides Cloudflare credentials:

npm run mcp:deploy

License

MIT