LLM Eval Harness
Run a golden set and gate LLM model swaps on accuracy, with latency and cost per request.
Description as published in the official MCP registry.
Maintain this server? Get the badge
What we measured
| Endpoint | None listed (runs locally) | registry |
|---|---|---|
| GitHub stars | 0 | github.com/razilzul/llm-eval-harness |
| Last push | 2026-10-18 | GitHub API |
| Licence | MIT | GitHub API |
| Registry version | 1.0.0 · active · updated 2026-10-18 | registry |
| Package | oci: ghcr.io/razilzul/llm-eval-harness:1.0.0 | registry |
Signed reports
No signed reports yet. A report carries a signed decision record from the reporter's gate, so it shows that the gate allowed a call, not just an opinion; it does not prove which server answered. How to file one.
Related servers
- Harness Start Claude Code and Codex agents the person can watch, wait for them and read their…
- agent-harnesses Agent-harness picks, decision guides, setup templates, and skills; live GitHub/HN…
- SandBase Harness Stdio MCP bridge for SandBase Harness agents, sessions, turns, artifacts, and…
- Palimpsest — censorship, China economy and model-eval observatory Rights-aware censorship, China-economic status, and tamper-evident AI evaluation tools.
- Barevalue AI podcast editing. Sign up in one call, send audio by link or from disk, get an edited…
- Sevalla Official Sevalla MCP — full PaaS API access through just 2 tools.
Is this your server?
Show the result of our latest check where your users look first. The badge is a live image: it changes when our check does, and it links back to this page.
[](https://openforallofus.com/tools/io.github.razilzul/llm-eval-harness)<a href="https://openforallofus.com/tools/io.github.razilzul/llm-eval-harness"><img src="https://openforallofus.com/badge/io.github.razilzul/llm-eval-harness.svg" alt="Agora: MCP handshake check" height="20"></a>The badge reports one thing: the result of our last handshake check. A server with no remote endpoint shows as "runs locally". How we check.
Agents can read this page as data: MCP endpoint https://openforallofus.com/api/mcp, tool get_tool with name io.github.razilzul/llm-eval-harness.