mcp-service
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., "@mcp-servicesearch the web for AI trends"
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.
mcp-service
Shared tool + skill host with two faces:
Standard MCP at
/mcp/(Streamable-HTTP, officialmcpSDK). Any MCP client connects and getstools/list·tools/call·prompts/list·prompts/get. Tools → MCP tools; skills → MCP prompts. This is the reusable, interoperable surface other tools/agents connect to.Management REST — CRUD for tools + skills (the job2cool admin UI), plus
POST /tools/{name}/invokeandGET /tools/manifest. (MCP is a consumption protocol, not authoring — so authoring lives here.)
Model
Tools are declarative catalog entries (name, JSON-schema, which built-in
impl, config, enabled). Implementations ship with the service (app/toolimpl.py); the UI never authors executable code (secure by design).Skills are reusable instruction templates (content + triggers). Exposed as MCP prompts. Trigger-based auto-injection is a consumer concern.
Related MCP server: SkillForge MCP
Networks
noted-network— apps (job2cool-backend) reach mcp-service.mcp_internal— mcp-service reaches the isolatedwebsearch_server.web_searchdelegates towebsearch_server /search.
Admin area (read-only)
Browser view over the whole registry (all apps at once) at /admin — search,
filter by app, toggle enabled-only, expand each tool's input_schema/config.
Backed by GET /admin/registry (aggregate JSON). Consult-only; authoring stays on
the CRUD REST face below.
REST quick-ref (default ?app=job2cool)
GET /health
GET /admin GET /admin/registry (read-only browser view)
GET /tools GET /tools/manifest
PUT /tools/{name} DELETE /tools/{name}
POST /tools/{name}/invoke {"args": {...}}
GET /skills PUT /skills/{name} DELETE /skills/{name}Run
docker network create mcp_internal # shared with websearch_server (once)
docker compose up -d --build
curl -s localhost:4950/health
curl -s -XPOST localhost:4950/tools/web_search/invoke \
-H 'content-type: application/json' -d '{"args":{"query":"kubernetes operator pattern","max_results":3}}'This server cannot be deployed
Maintenance
Related MCP Connectors
- UnifAPIOAuthcom.unifapi
Hosted MCP server for live public-data APIs and Skills for AI agents.
Remote MCP server exposing SMI Aware tools, resources, and skills over Streamable HTTP.
MCP server for progressive tool usage at any scale (see https://klavis.ai)
Personal assistant MCP server with search, execute, packages, jobs, secrets, and integrations.
Related MCP Servers
- AlicenseNot gradedqualityNot gradedmaintenanceAn MCP server that transforms Claude-style skills and resources into callable tools for any MCP-compatible agent or client. It automatically discovers, exposes, and executes scripts from skills organized in local directories or packaged archives.MIT
- AlicenseBqualityBmaintenanceA universal MCP server that loads Markdown skills from arbitrary folders, enabling lazy discovery and execution of skills across multiple MCP-capable tools.517 npm2MIT
- AlicenseNot gradedqualityBmaintenanceA self-hostable MCP server that turns a folder of skills into callable tools via MCP and REST APIs.2MIT
- AlicenseNot gradedqualityBmaintenanceAn MCP server that makes Agent Skills available to any MCP-compatible agent through a declarative, package.json-based configuration, enabling team-shareable skill management and execution.4MIT