Standalone MCP proxy service — manages persistent connections to external MCP servers
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"mcp-plugin","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# slife-mcp Standalone MCP proxy service — manages persistent connections to external MCP servers and exposes them through a single MCP endpoint. ## Quick Start ```bash pip install slife-mcp # Run — auto-detects mode: slife-mcp # with slife.json5 → HTTP, without → stdio slife-mcp --port 8888 # HTTP on custom port slife-mcp --host 0.0.0.0 --port 9876 # HTTP on all interfaces ``` ## How It Works ``` MCP clients ←→ slife-mcp ←→ external MCP servers (filesystem, search, fetch, …) ``` slife-mcp acts as a **proxy layer** between MCP clients and external MCP servers. Instead of each client managing its own server connections, slife-mcp maintains persistent connections centrally and exposes all tools through a single endpoint. Key design decisions: - **Raw JSON-RPC over subprocess pipes** — no `anyio`, no `ClientSession`. Avoids `TaskGroup` conflicts with FastMCP and keeps the implementation simple and debuggable. - **Connection pooling** — servers are kept alive across requests. Disconnected servers can be reconnected at runtime without restarting slife-mcp. - **Stderr capture** — when a server fails to connect, its stderr output is included in the error message so the LLM (or human operator) can understand and fix the problem. - **Progressive disclosure** — servers default to eager mode (tools loaded at startup), but can be configured as lazy (connected, tools hidden) to keep the tool list lean. Switch between modes at runtime with `mcp_set_disclosure`. This supports any MCP-compatible server, including: - **Pre-built MCP servers** — filesystem, web search (Serper, Tavily), fetch, etc. - **REST API servers** — [anyapi-mcp-server](https://github.com/quiloos39/anyapi-mcp-server) converts OpenAPI specs to MCP tools, making any REST API (GitHub, Jira, GitLab, Slack…) callable as tools. ## Configuration Create `slife.json5` in your working directory: ```json5 { mcp: { // wrapper.url — where clients connect to slife-mcp wr…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.