io.theagentforge/forge-gauntlet

up checked 2026-11-11T10:38:44.444Z

Unclaimed · Unverified

Publisher-provided title

The Forge Gauntlet

Publisher-provided description

See whether your own agent falls for the traps AI models fall for: six refereed jobs. No account.

Summary by lookupd

Runs an agent through refereed sandbox jobs with planted traps and reports whether it avoids them, without requiring an account.

Install

Claude Code, remote (streamable-http)

claude mcp add --transport http forge-gauntlet https://theagentforge.io/mcp/gauntlet

Codex, remote (streamable-http)

codex mcp add forge-gauntlet --url https://theagentforge.io/mcp/gauntlet

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

{
  "mcpServers": {
    "forge-gauntlet": {
      "url": "https://theagentforge.io/mcp/gauntlet"
    }
  }
}

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

Publisher-provided text

gauntlet_start Use this when the user wants their own agent tested on everyday jobs with planted traps, refereed from the end state. Inputs: optional mode ("sprint": one job,
gauntlet_act Use this during a Gauntlet run to call one sandbox tool in the current job. Inputs: gauntlet_id, tool and args. Returns: the tool's result; when a job ends, its
gauntlet_status Use this to see where a Gauntlet run stands. Inputs: gauntlet_id. Returns: the current job (brief, tools, calls left), or the report card once finished. Limits:

Shows this live status and links here; updates hourly.

[![lookupd status](https://lookupd.com/badge/io.theagentforge/forge-gauntlet.svg)](https://lookupd.com/servers/io.theagentforge/forge-gauntlet)