io.github.saagpatel/operant-mcp

up checked 2026-11-11T05:39:18.261Z

Unclaimed · Unverified

Publisher-provided title

operant-mcp

Publisher-provided description

Read-only MCP server for the OPERANT AI operating-agent calibration benchmark.

Summary by lookupd

Read-only access to benchmark results, case metadata, and methodology for comparing AI model calibration scores.

Install

Claude Code, remote (streamable-http)

claude mcp add --transport http operant https://operant-mcp.saagar210.workers.dev/mcp

Claude Code, npm package (stdio)

claude mcp add operant -- npx -y saagar-operant-mcp

Codex, remote (streamable-http)

codex mcp add operant --url https://operant-mcp.saagar210.workers.dev/mcp

Codex, npm package (stdio)

codex mcp add operant -- npx -y saagar-operant-mcp

Cursor, remote (streamable-http) → .cursor/mcp.json

{
  "mcpServers": {
    "operant": {
      "url": "https://operant-mcp.saagar210.workers.dev/mcp"
    }
  }
}

Cursor, npm package (stdio) → .cursor/mcp.json

{
  "mcpServers": {
    "operant": {
      "command": "npx",
      "args": [
        "-y",
        "saagar-operant-mcp"
      ]
    }
  }
}

Other clients: docs. Or tell your agent: Set up lookupd for me: read https://lookupd.com/agent.md and follow it.

Publisher-provided text

get_results Return retained model calculation profiles plus generated_at, claim_status, claims_at_risk, and the bounded evidence_binding. These rows are not durable named-m
compare_models Place two retained calculation profiles side by side by display_name substring. Returns ocs_mean, ocs_stdev, orchestration_mean, run_family, and subject_shell f
get_methodology Return the benchmark design: the 4 axes, the OCS formula (Youden's J), the 5 decision labels (PROCEED, PROCEED_SANCTIONED, REFUSE, ESCALATE, REROUTE), scoring b
list_cases Return case metadata (no full task prompts): pair_id/id, axis, tier, grounding, and side indicators (malign/benign for axes 1/2/4; null for axis 3). Filter by a
get_case Return the full case for a given pair_id (axes 1/2/4) or id (axis 3): malign and benign task prompts, expected decisions, grounding rationale, and bypass patter

Shows this live status and links here; updates hourly.

[![lookupd status](https://lookupd.com/badge/io.github.saagpatel/operant-mcp.svg)](https://lookupd.com/servers/io.github.saagpatel/operant-mcp)