LLMPvP: an MCP arena where LLM agents duel in ranked Chess and Go with Glicko-2
vudueprajacu · reddit · 2026-09-06
A developer released LLMPvP, a bring-your-own-LLM MCP server where AI agents play ranked Chess and Go against each other with separate Glicko-2 ratings per game. The model runs on your side—your LLM API key never touches LLMPvP's servers.
The npm package llmpvp-plugin exposes 9 tools: agent registration/status, matchmaking, challenges (or house bots), game state, moves, and resignation. It's on the official MCP Registry and runs via npx llmpvp-plugin mcp.
MCP design lessons shared:
- The 60-second move timeout is stated directly in the tool description—letting the model know before moving beats erroring after the fact.
- Illegal moves and timeouts share a 3-strike limit, so one bad tool call doesn't end the game; after 3 strikes the game is lost "by conduct."
More from coding & agent
- Spec-to-APK entirely on an Android phone via Termux CLI and AI — Digital_Otorongo · 2026-09-06
- Astra review: executes hour-long agentic tasks while iterating on the plan without losing the thread — ivan_bezdomny · 2026-09-06
- Stop picking coding models on vibes: run your own A/B/C refactor tests — robleclerc · 2026-09-06
- Orca, the 62k-star ADE for running parallel coding agents, recommended for Mac Mini dev setups — kieranklaassen · 2026-09-06
- Dev Abandons LLM Agent Harness After 883 Commits and 8 Months, Open-Sources the Wreckage — dancingwithlies · 2026-09-06
- Dev one-shots a UEVR fork with GPT 6 Astra for true stereoscopic VR windows — pvncher · 2026-09-06