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.
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 rootskills/directory.
Pull from the public docs endpoint:
npm run sync:skillsPull from a local docs checkout:
WAVEDASH_SKILLS_SOURCE_DIR=/path/to/wvdsh/docs/skills npm run sync:skillsGitHub Actions runs the same sync daily and commits changes when docs updates change the generated skill files.
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 startThe endpoint is http://localhost:3000/mcp.
Validate the Cloudflare Worker bundle without publishing:
npm run mcp:deploy:dry-runDeploy the MCP Worker from an environment that provides Cloudflare credentials:
npm run mcp:deployMIT