# MCP Selection Lab

> Benchmark MCP tool selection with metadata-only routing, collision, abstention, and holdout checks.

- name: io.github.ChanghuLiu/mcp-selection-lab
- version: 0.5.2
- connection class: R0 (autonomous)
- trust: 97/100
- quality: 96/100 (strong)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/io.github.ChanghuLiu%2Fmcp-selection-lab
- tools: https://api.protogrid.dev/v1/servers/io.github.ChanghuLiu%2Fmcp-selection-lab/tools

**An agent can connect right now, no human step.** Remote endpoint, no authentication, reachable on the last probe.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "mcp-selection-lab": {
      "type": "http",
      "url": "https://agenttrustlab.com/mcp"
    }
  }
}
```


## Trust

- hygiene: 100
- liveness: 100
- freshness: 100
- provenance: 90
- drivers: +repository +namespace-matches-repo +website +reachable +uptime-100% +updated-7d-ago
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 96/100 (strong)

### protocol: 88 (weight 25)

- pass `protocol.modern`: supports 2026-07-28 (server/discover)
- pass `protocol.stateless`: answers without a session
- pass `protocol.transport`: streamable HTTP
- warn `protocol.list_ttl`: tools/list declares ttlMs 0, so clients re-fetch it on every use

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 100 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- pass `tools.annotations`: 3 of 3 tools declare readOnlyHint or destructiveHint
- pass `tools.directory_hints`: every tool declares readOnlyHint, destructiveHint, idempotentHint and openWorldHint
- pass `tools.token_cost`: about 464 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 100 (weight 15)

- pass `stability.changes`: no tool changes in 30 days
- pass `stability.rug_pull`: no tool changed its meaning under the same name

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 3, about 464 tokens to load them all
- tool set hash: 8ffc54c1ad57939c90147547675bc892
- badge: [![protogrid quality](https://protogrid.dev/badge/io.github.ChanghuLiu/mcp-selection-lab.svg)](https://protogrid.dev/servers/io.github.ChanghuLiu/mcp-selection-lab)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

_No tool definition changes recorded._

## Remotes

- https://agenttrustlab.com/mcp (streamable-http, auth none, reachable true, uptime30d 1)

## Packages

_none_

## Tools (3)

- `get_selection_report`: Read and agent-enrich an existing public scan report by id; no new target request or model call.
- `scan_mcp_metadata`: Create a free public report of MCP tool-description routing collisions.

    Reads discovery metadata (initialize/tools/list) only; never calls target business
    tools or spends funds. Use a public HTTP(S) URL without credentials, query or fragment.
    Results use at most 12 tools and 24 generated cases, not a real-model evaluation.
    The returned report is enriched with agent_plan candidate description-only edits and
    a machine-readable rerun instruction. Candidate edits are not untouched-holdout proof.
    Reports persist and are public by link. For an existing report use get_selection_report.
    Set internal_test=true for owner/CI validation so it is excluded from public scans.
    source is a self-reported referral bucket, not identity verification.
    
- `selection_lab_info`: Get free service capabilities, limits, privacy and connection details; does not scan a target.