Skip to main content
Glama
authensor-mcp-server

authensor-mcp-server

Official
by AUTHENSOR

Authensor is a security research firm in San Francisco, founded by John Daniel Kearney. Most repositories on this profile are forks kept for upstream pull requests; the work lives in etb-scan, etb-harness and the repositories listed below. It audits the instruments the AI industry measures itself with: evaluation harnesses, benchmarks, LLM judges, leaderboards and scoring pipelines.

Every defect we report is proven with a running reproduction against the genuine code at a pinned commit, and disclosed to the maintainer first. Fixed-scope audits end in an attestation with a scope, a standard version and an expiry date.

Research

  • The Evaluator Trust Boundary (ETB-01 to ETB-10). An evaluator has an ETB failure when its score depends on an artifact the evaluated system can influence. The page lists dated cases with upstream fixes in SWE-bench, Promptfoo, NVIDIA garak and the UK AI Security Institute's Inspect and ControlArena. Preprint: 10.5281/zenodo.21633620.

  • etb-scan, an offline scanner that measures ETB-01 (injected verdicts) against your own judge. Standard library only, no API keys, MIT.

pip install etb-scan
etbscan --judge mypkg:judge --trials 10
  • etb-harness, which measures how gameable an agent benchmark is before and after hardening: deterministic capability probes (can the agent read the tests, reach the fix commit, forge a pass), opt-in LLM attacker policies, a verifier and hardening profiles. First targets: SWE-bench Verified, ControlArena. Apache-2.0. Results are published only after the maintainers have been notified.

  • The ETB registry, one OSV-format record per confirmed evaluation-integrity flaw (ETB-YYYY-NNNN), with the disclosure policy and the full taxonomy (ETB-01 to ETB-10 published, ETB-11 to ETB-16 proposed). Opens with the first published entries.

  • The ledger: every defect filed, every fix that landed upstream, and every target read and found sound. Includes the MLflow CVE-2026-64849 credit.

Related MCP server: evav-gateway

Work with us

Evaluation-integrity audits, and a free claim check: one published number checked against the instrument that produced it, within 48 hours.

authensor.com · john@authensor.com


The retired agent-safety toolkit

This repository first held Authensor's open-source agent-safety toolkit: the policy engine, the Aegis content scanner, the Sentinel monitor, SafeClaw and the @authensor npm packages. That product line is retired and no longer maintained. Its code and documentation stay here for reference:

Quickstart · Features · Red teaming · API reference · Deployment · Development

License

MIT.

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
      D
      maintenance
      A policy-enforcing MCP gateway that intercepts all tool calls to downstream MCP servers, applying allow/deny/ask rules with human approval and audit logging for safe access to dangerous tools.
      7 npm
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Governed MCP gateway that lets AI agents call tools with policy enforcement, prompt-injection screening, a kill-switch, and tamper-evident signed audit logs.
      Apache 2.0
    • A
      license
      Not graded
      quality
      B
      maintenance
      Provides a governance layer for MCP servers, enforcing policy-based allow/deny/approval rules, requiring human approval for risky calls, and recording an auditable trail of every AI agent tool call.
      MIT
    • A
      license
      Not graded
      quality
      B
      maintenance
      Enforces allow/deny/ask policies on MCP tool calls before they reach downstream servers, with tamper-evident audit logging.
      MIT