Bench Agent Discovery MCP Server

Discover public AI agents, reusable recipes, and trusted benchmark evidence by task.

Local serverstdioPython

What is the Bench Agent Discovery MCP server?

Discover public AI agents, reusable recipes, and trusted benchmark evidence by task. That is what the bench agent discovery mcp server brings to an AI assistant: the same capability, reachable through the Model Context Protocol rather than a separate app or dashboard.

The short version

One line of code. Live dashboard, public profile, README badge.

You built an AI agent. You ran it a few times. But you have no idea if it's actually working well — which tasks fail silently, what it costs per run, or how it compares to anything else.

The tools it exposes

The server publishes 8 tools. What each one is for:

  • Feature — Description
  • Leaderboard — Browse public agents by runs, success rate, eval score, or cost
  • Compare — /vs/@a/agent1/@b/agent2 — side-by-side quality, cost, latency
  • Benchmarks — Versioned benchmark suites with repeated runs and evidence trails. Separate from self-reported telemetry
  • Product — Role
  • Workers — API, profile rendering, badge generation
  • Workflows — Durable retry for EvalWorkflow and ClusterWorkflow
  • Assets — Static frontend (landing, dashboard, JS, CSS)

What it needs from you

Configuration is passed through the environment: BENCH_KEY, OTEL_EXPORTER_OTLP_ENDPOINT, OTEL_EXPORTER_OTLP_PROTOCOL, OTEL_EXPORTER_OTLP_HEADERS, SESSION_SECRET, GITHUB_OAUTH_CLIENT_SECRET. Treat anything key-shaped as a real credential — scope it to the minimum the server needs, and rotate it if it ever lands in a shared config.

Getting it running

@virajmishra1/bench-cli on npm is all you need. Most clients run it directly, so configuration is a few lines and a restart.

How it compares

This sits in the planning and project tracking group, where several servers overlap in what they claim to do but differ sharply once you actually set them up. Bench Agent Discovery's toolset — Feature, Leaderboard, Compare and 5 more — is a fair guide to whether it matches your workflow. It is maintained by VirajMishra1; worth a glance at recent repository activity before you build anything load-bearing on it.

We check each listing at SyncDev against the project's documentation before it goes live — if something here drifts out of date, it is a bug worth reporting.

Things to watch

  • It runs with your machine's permissions. That is convenient and also the reason to think about what you point it at before you approve a tool call.
  • Missing credentials fail quietly in some clients — if no tools show up, check the environment block first.
  • Keep per-call confirmation enabled while you learn its behaviour; it is the cheapest safeguard you have.

Available tools

ToolWhat it does
FeatureDescription
LeaderboardBrowse public agents by runs, success rate, eval score, or cost
Compare/vs/@a/agent1/@b/agent2 — side-by-side quality, cost, latency
BenchmarksVersioned benchmark suites with repeated runs and evidence trails. Separate from self-reported telemetry.
ProductRole
WorkersAPI, profile rendering, badge generation
WorkflowsDurable retry for EvalWorkflow and ClusterWorkflow
AssetsStatic frontend (landing, dashboard, JS, CSS)

How to install the Bench Agent Discovery MCP server

{
  "mcpServers": {
    "bench": {
      "command": "npx",
      "args": ["-y", "@virajmishra1/bench-cli"],
      "env": {
        "BENCH_KEY": "your-value",
        "OTEL_EXPORTER_OTLP_ENDPOINT": "your-value",
        "OTEL_EXPORTER_OTLP_PROTOCOL": "your-value",
        "OTEL_EXPORTER_OTLP_HEADERS": "your-value",
        "SESSION_SECRET": "your-value",
        "GITHUB_OAUTH_CLIENT_SECRET": "your-value"
      }
    }
  }
}

Add to claude_desktop_config.json, then restart Claude Desktop.

Configuration

VariableDescriptionRequired
BENCH_KEYCredential the server authenticates with.Yes
OTEL_EXPORTER_OTLP_ENDPOINTConfiguration value read at startup.Optional
OTEL_EXPORTER_OTLP_PROTOCOLConfiguration value read at startup.Optional
OTEL_EXPORTER_OTLP_HEADERSConfiguration value read at startup.Optional
SESSION_SECRETCredential the server authenticates with.Yes
GITHUB_OAUTH_CLIENT_SECRETCredential the server authenticates with.Yes

Example prompts to try

  • Use Bench Agent Discovery to Feature.
  • Use Bench Agent Discovery to Leaderboard.
  • Use Bench Agent Discovery to Compare.

Frequently asked questions

It connects Bench Agent Discovery to MCP-compatible AI assistants such as Claude and Cursor, exposing 8 tools (Feature, Leaderboard, Compare, and more) that the assistant can call on your behalf. Instead of copying data back and forth by hand, the assistant works with Bench Agent Discovery directly.