Developer Infrastructure · Product Profile v2
MCP Selection Lab
A benchmark and scan toolkit for investigating how agents choose among MCP tools.
Evidence Tested
What is it?
A benchmark and scan toolkit for investigating how agents choose among MCP tools.
| Provider / organization | Changhu Liu |
|---|---|
| Homepage / project | https://mcp-selection-lab-production.up.railway.app/ |
| Repository | https://github.com/ChanghuLiu/mcp-selection-lab |
| MCP endpoint | https://mcp-selection-lab-production.up.railway.app/mcp |
| Authentication | NONE for documented public discovery tools |
| Transport | streamable-http |
| Tool count | 3 |
| Last verified | 2026-09-12 |
What evidence do we have?
Evidence state: EVIDENCE_TESTED
Existing S2 selection evidence and Reliability Explorer records remain canonical; this profile does not rescore them.
What can an agent use it for?
- Run bounded metadata-only selection-readiness scans.
- Read existing selection evidence and reports.
What type of tool surface does it expose?
Three public discovery/report tools; no target business-tool execution.
What remains untested or unknown?
- Current behavior may differ from the cited source snapshot or documentation.
Source provenance
Has Agent Trust Lab tested it?
Yes — existing evidence is linked above.