Koskamo

Bench Agent Discovery MCP

A public read-only MCP server that allows agents to discover public AI agents, view reusable recipes, and access trusted benchmark evidence by task. It provides tools to search for agents, retrieve specific agent details, and list available benchmarks, with owner telemetry and benchmark evidence labeled separately.

Endpoint and access

No sign-in to connect and list tools. Individual tools may still need a vendor account, credentials, or permissions; check the documented prerequisites.

Bench Agent Discovery MCP logo, by VirajMishra1Community listingStreamable HTTPNo sign-inEndpoint checked 2026-11-06Published by VirajMishra1
MCP endpoint (Streamable HTTP)
https://bench.virajmishratakehome.workers.dev/mcp
Vendor documentation

What its tools do

3 tools, grouped by the hints the server advertises. Hints are self-declared, not verified behavior.

Read-only (3)

3 of 3 have a description · 3 of 3 declare behavior hints · 1 of 2 with parameters document them all

Checks and signals

Each line is separate evidence, not a trust score.

  • Passed Endpoint answered the live check ·
  • Availability 1 of 1 checks passed in 30 days

What you can do with it

  • Search for public AI agents
  • Get details for a specific agent
  • List available benchmark suites

Source and maintenance

Community listingRegistry 0.1.0
mcp-registry: io.github.VirajMishra1/bench@0.1.0Source repositoryProject website
Maintenance checked

Protocol and capabilities

Server
Bench Agent Discovery 0.1.0
MCP protocol negotiated in check
2025-11-25
Tools
3 advertised
Resources
Not advertised
Resource templates
Not advertised
Prompts
Not advertised
Server-advertised capabilities
{
  "tools": {
    "listChanged": true
  }
}

Server-provided instructions

Untrusted server output shown for inspection. It is not Koskamo guidance and is excluded from knowledge search.

Use search_agents to find listed public agents by task, then get_agent for evidence and reuse details. Treat owner_telemetry as observed self-submitted workloads, never as a controlled ranking. Treat only verified_benchmarks as controlled evidence. This server is read-only; invocation is not available here.

Tools

3 advertised

search_agents — Search public AI agents
Advertised by the server during the last successful anonymous check.
Advertised read-only

Find listed public agents by task, capability, category, framework, model, verified evidence, or reuse configuration. Owner telemetry and controlled benchmark evidence are returned separately.

search_agents input schema
{
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "properties": {
    "sort": {
      "enum": [
        "verified",
        "recent",
        "runs"
      ],
      "type": "string",
      "default": "verified"
    },
    "limit": {
      "type": "integer",
      "default": 10,
      "maximum": 20,
      "minimum": 1
    },
    "model": {
      "type": "string",
      "maxLength": 80
    },
    "query": {
      "type": "string",
      "maxLength": 120,
      "description": "Task or capability to search for, such as grounded research or code review."
    },
    "license": {
      "type": "string",
      "maxLength": 40,
      "description": "Exact SPDX-style license id from the agent's manifest provenance, such as MIT or Apache-2.0."
    },
    "category": {
      "type": "string",
      "maxLength": 60
    },
    "reusable": {
      "type": "boolean",
      "description": "True returns agents whose owners configured an invocation policy and capability manifest."
    },
    "verified": {
      "type": "boolean",
      "description": "True returns agents with at least one trusted-runner-verified benchmark submission."
    },
    "framework": {
      "type": "string",
      "maxLength": 60
    },
    "liveCallable": {
      "type": "boolean",
      "description": "True returns agents with a reusable invocation policy and an owner-verified, currently reachable endpoint."
    },
    "maxP50LatencyMs": {
      "type": "integer",
      "minimum": 0,
      "description": "Upper bound on the agent's observed p50 latency in milliseconds."
    },
    "maxCostPerRunUsd": {
      "type": "number",
      "minimum": 0,
      "description": "Upper bound on lifetime total_cost_usd / total_runs, i.e. average observed cost per run."
    }
  },
  "additionalProperties": false
}
search_agents annotations
{
  "readOnlyHint": true,
  "openWorldHint": false,
  "idempotentHint": true,
  "destructiveHint": false
}
get_agent — Get a public agent
Advertised by the server during the last successful anonymous check.
Advertised read-only

Get one public agent's recipe, public capability manifest, coarse invocation status, owner telemetry, and verified benchmark submissions.

get_agent input schema
{
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "required": [
    "handle"
  ],
  "properties": {
    "handle": {
      "type": "string",
      "maxLength": 101,
      "description": "Bench handle in @owner/agent-slug form."
    }
  },
  "additionalProperties": false
}
get_agent annotations
{
  "readOnlyHint": true,
  "openWorldHint": false,
  "idempotentHint": true,
  "destructiveHint": false
}
list_benchmarks — List verified benchmark contracts
Advertised by the server during the last successful anonymous check.
Advertised read-only

List public, versioned benchmark contracts and only their trusted-runner-verified submissions.

list_benchmarks input schema
{
  "type": "object",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "properties": {}
}
list_benchmarks annotations
{
  "readOnlyHint": true,
  "openWorldHint": false,
  "idempotentHint": true,
  "destructiveHint": false
}

Prompts

Not advertised

No prompt definitions available.

Resources

Not advertised

No resource definitions available.

90-day uptime

A lightweight connection check about every six hours, sampled on hosts with many listings. It shows the server answered, not that its tools work.

Too few checks for a percentage yet (2 of 2 answered).

  1. 2026-11-06: 1 of 1 checks answered
  2. 2026-11-07: 1 of 1 checks answered

30-day availability

Aggregated from read-only remote checks; this is not a tool-execution or security guarantee.

Successful checks
1 of 1 (100%)
Average latency
262 ms
95th percentile latency
262 ms

Maintenance history

Public check and source-refresh events. Editorial and ownership records remain private.

  • · probe · passed · 262 ms · metadata changed · 3 tool(s)

Connect from your client

Each snippet registers the same endpoint; pick the client you use.

This server is operated by the named third party. Listing and endpoint checks are not an endorsement or a security audit. Review the operator's permissions, terms, and data handling before connecting. How listings are checked and corrected.

Claude Code

Run this once in a terminal. It registers the remote server. Use --scope project to write .mcp.json for the repo.
Claude Code configuration
claude mcp add --transport http bench-agent-discovery-mcp https://bench.virajmishratakehome.workers.dev/mcp

Codex

Run this once. Codex stores MCP servers in ~/.codex/config.toml (or .codex/config.toml in a trusted project).
Codex configuration
codex mcp add bench-agent-discovery-mcp --url https://bench.virajmishratakehome.workers.dev/mcp

Cursor

Merge this server into .cursor/mcp.json in your project, then enable bench-agent-discovery-mcp in your MCP settings.
Cursor configuration
{
  "mcpServers": {
    "bench-agent-discovery-mcp": {
      "url": "https://bench.virajmishratakehome.workers.dev/mcp"
    }
  }
}

Cline

Merge this into cline_mcp_settings.json (Cline → MCP Servers → Configure).
Cline configuration
{
  "mcpServers": {
    "bench-agent-discovery-mcp": {
      "type": "streamableHttp",
      "url": "https://bench.virajmishratakehome.workers.dev/mcp",
      "disabled": false
    }
  }
}

VS Code

Merge this server into .vscode/mcp.json. Use MCP: List Servers to start bench-agent-discovery-mcp and inspect its tools.
VS Code configuration
{
  "servers": {
    "bench-agent-discovery-mcp": {
      "type": "http",
      "url": "https://bench.virajmishratakehome.workers.dev/mcp"
    }
  }
}

Gemini CLI

Run this once. It writes .gemini/settings.json in the current project; pass --scope user for ~/.gemini/settings.json.
Gemini CLI configuration
gemini mcp add --transport http bench-agent-discovery-mcp https://bench.virajmishratakehome.workers.dev/mcp

Claude Desktop

Open Claude Settings > Connectors > Add custom connector, enter this server URL and enable it for the conversation. This is a manual remote-connector handoff, not a desktop JSON configuration.
Claude Desktop configuration
https://bench.virajmishratakehome.workers.dev/mcp

Windsurf

Merge this into ~/.codeium/windsurf/mcp_config.json, then refresh MCP in Cascade. Windsurf remote servers use serverUrl, not url.
Windsurf configuration
{
  "mcpServers": {
    "bench-agent-discovery-mcp": {
      "serverUrl": "https://bench.virajmishratakehome.workers.dev/mcp"
    }
  }
}

Let your agent find its tools

Find servers like Bench Agent Discovery MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.

Connect Koskamo Explore workflows
See all 5,557