evals MCP servers

25 tagged evals, measured the same way as everything else here.

Browse within: LLM 7Multi-Agent 5model-context-protocol 5nextjs 5

runtime_proof

01

harbor-framework/harbor

MCP server Claude CodeCodexCursor +2

MCP server "runtimeproof", hosted remotely at runtime-mcp-server, as configured in harbor-framework/harbor.

not rated 6.2k +197 today A tokens not measured original Apache-2.0

scenario-assistant

02

prime-radiant-inc/scenarios

MCP server Claude CodeCodexCursor +2

MCP server "scenario-assistant" as configured in prime-radiant-inc/scenarios. Runs locally from the ../../tools/mcp_server.py Python package.

not rated 10 4mo ago A tokens not measured original Apache-2.0

commutescout

03

nicglazkov/commutescout

MCP server Claude CodeCodexCursor +2

Live California road conditions: CHP incidents, Caltrans closures, chain controls, and wildfires. Remote server at mcp.commutescout.com.

not rated 4 4d ago A tokens not measured original MIT

opengate-mcp

04

nickjlamb/opengate

MCP server Claude CodeCodexCursor +2

Check whether an AI answer is grounded in its context — deterministic, no LLM judge. Runs locally from the @pharmatools/opengate-mcp npm package.

not rated 1 +1 18d ago A tokens not measured original MIT

querypilot

05

nickklos10/QueryPilot

MCP server Claude CodeCodexCursor +2

Eval-driven SQL reliability for AI agents. Runs locally from the querypilot Python package. Needs 6 environment variables to run.

not rated 1 3mo ago A tokens not measured original MIT

canopy

06

kruslim/canopy

MCP server Claude CodeCodexCursor +2

Python Client for the Canopy Simulation API. Runs locally from the canopy Python package.

not rated 1 +1 3mo ago A tokens not measured

rule-drift

07

ri7in/rule-drift

MCP server Claude CodeCodexCursor +2

Test whether AI agents retain critical instructions as conversations grow. Runs locally from the rule-drift npm package.

not rated 0 +1 3d ago A tokens not measured original MIT

safetyreplay-mcp

08

clauxel/safety-replay-gate-mcp

MCP server Claude CodeCodexCursor +2

Paid remote MCP for safety replays, policy gates, eval receipts, and release evidence. Remote server at safetyreplay.clauxel.com.

not rated 0 5mo ago A tokens not measured

auxiliar-mcp

09

auxiliar-ai/auxiliar-mcp

MCP server Claude CodeCodexCursor +2

Eval-backed discovery for the auxiliar.ai gateway — the best web-access provider per job, measured. Runs locally from the auxiliar-mcp npm package.

not rated 0 3mo ago A tokens not measured original MIT archived

redline

10

gowtham0992/redline

MCP server Claude CodeCodexCursor +2

Generate eval suites from prompt logs and catch LLM regressions locally. Runs locally from the redline-ai Python package.

not rated 0 2mo ago A tokens not measured original MIT

ai-workbench-mcp

11

hrishikesh-thakre/ai-workbench-mcp

MCP server Claude CodeCodexCursor +2

Goose-first acceptance, validation, routing, and audit layer for agentic work. Runs locally from the ai-workbench-mcp Python package.

not rated 0 3mo ago A tokens not measured original Apache-2.0

multivon-mcp

12

multivon-ai/multivon-mcp

MCP server Claude CodeCodexCursor +2

MCP server exposing multivon-eval + pdfhell as agent-callable tools. Drop into Claude Desktop, Cursor, Cline, or any MCP-compatible AI coding agent. Runs locally from the multivon-mcp Python package.

not rated 0 1mo ago A tokens not measured

pantry-live

13

ketankshukla/learn-mcp-agent-crew

MCP server Claude CodeCodexCursor +2

One of 5 in .mcp.json

MCP server "pantry-live", hosted remotely at learn-mcp-agent-crew.vercel.app, as configured in ketankshukla/learn-mcp-agent-crew.

not rated 0 1mo ago A tokens not measured

grounded-mcp

14

rich-atkins/grounded-mcp

MCP server Claude CodeCodexCursor +2

MCP server for real-time package versions, download counts, and service status. Covers 1,000+ npm/PyPI packages and 76 cloud services. Runs locally from the grounded-mcp Python package.

not rated 0 23d ago A tokens not measured original MIT

agent-eval-gate

15

linus10x/agent-eval-gate

MCP server Claude CodeCodexCursor +2

MCP server "agent-eval-gate" as configured in linus10x/agent-eval-gate. Runs locally from the agent-eval-gate Python package.

not rated 0 13d ago A tokens not measured original MIT

rigforge

16

mrodgersjs-web/rigforge

MCP server Claude CodeCodexCursor +2

MCP server "rigforge" as configured in mrodgersjs-web/rigforge. Runs locally from the rigforge Python package.

not rated 0 6d ago A tokens not measured original MIT

foreman

17

YashNirwan/foreman

MCP server Claude CodeCodexCursor +2

Python library to manipulate The Foreman through the API. Runs locally from the foreman Python package.

not rated 0 13d ago A tokens not measured original MIT

pipsync-mcp-evals

18

pipsyncio/pipsync-mcp-evals

MCP server Claude CodeCodexCursor +2

MCP server "pipsync-mcp-evals" as configured in pipsyncio/pipsync-mcp-evals. Runs locally from the pipsync-mcp-evals Python package.

not rated 0 1mo ago A tokens not measured original Apache-2.0

oficio

19

brayans7/oficio

MCP server Claude CodeCodexCursor +2

A calm interface between an Obsidian daily note and Hermes Agent. Runs locally from the oficio Python package.

not rated 0 1mo ago A tokens not measured original MIT

claude-flow

20

amangit1314/repo-rag

MCP server Claude CodeCodexCursor +2

Ruflo CLI - Enterprise AI agent orchestration with 60+ specialized agents, swarm coordination, MCP server, self-learning hooks, and vector memory for Claude Code. Runs locally from the @claude-flow/cli npm package. Needs 6 environment variables to run.

not rated 0 21d ago A tokens not measured

At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: