Skip to main content
Glama

Collusion Sentinel for Agent Markets

Server Details

Emergent collusion leaves no agreement to find — it lives in the history.

Ownership verified
Status
Healthy
Uptime
100.0% over the last 24 hours
Last Tested
Transport
Streamable HTTP · MCP 2025-03-26
URL
ChatGPT
–
Claude
–

TDQS

C2.8/5.0

Scored across 3 tools

Disambiguation4/5

The three tools have mostly distinct roles: audit inspects histories, certify runs a certification battery, and control_plan produces a remediation plan. However, audit and certify both assess collusion risk, so an agent could initially hesitate over which to use without reading descriptions closely.

Naming Consistency4/5

All tools share the sentinel_ prefix and snake_case convention, making the set predictable. The only minor inconsistency is that audit/certify are verbs while control_plan is a noun phrase, but the pattern remains readable.

Tool Count5/5

Three tools fit this narrow, specialized purpose well: detect, certify, and remediate. Each tool earns its place without obvious filler or missing surface area.

Completeness4/5

The surface covers the core lifecycle of auditing, certifying, and planning remediation for collusion risk. Minor gaps exist around retrieving or verifying an issued certificate, but agents can likely work around this using the returned certificate hash.

Available Tools

3 tools
sentinel_auditBInspect

Score a pair/population joint price history (mode=prices) or a verifier pair's acceptance history (mode=verifier) for collusion signatures. $0.05/audit x402 after the one-time 100-audit grant; degenerate histories refused, never charged.

Inputs: rounds (required), mode.

sentinel_certifyBInspect

Run the tempt-to-violate certification battery for a subject ($25/certificate x402; scripted simulation battery over an operator-declared configuration — the subject's live model is not run). model and prompt_hash are required — the certificate hash binds them.

Inputs: kind (required), model (required), subject (required), prompt_hash (required).

sentinel_control_planDInspect

Score-keyed counterparty rotation + memory hygiene plan (included).

Inputs: audit_id (required), pair.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      C
      maintenance
      Multiplayer coordination for AI coding agents: Claude Code, Codex CLI and Cursor share one room per repository. An agent claims a path glob before it edits and a conflicting claim is refused at claim time, so collisions are prevented rather than resolved at merge. Metadata only — source code and diffs never leave the machine.
      MIT
    • A
      license
      B
      quality
      C
      maintenance
      An ultra-rational A2A protocol for zero-token edge pre-filtering and FEP-driven deadlock prevention. Uses Cloudflare Vectorize (384d cosine similarity) with a 24h deposit model, restricting bargaining to a 4-rally limit before forcing HTTP 402 dimension jumps.
      2
      MIT
    Try in Browser

    Glama MCP Gateway

    Add one secure layer between your agents and this server.

    Resources