agentmods.dev

glassbox-framework

A local MCP server for checking the reasoning and claims in AI answers. MCP is a standard way for an AI agent to use external tools, and this server runs through the npm package manager.

In plain words
What is it for?
Use it to run claim checks, constitutional checks, red-team reviews, and audits of AI answers from a local MCP setup.
Why use it?
It adds a local checking step for answers, including reviews and red-team checks. It requires the listed environment settings and an Anthropic API key.

MCP server for Claude CodeCodexCursorGemini CLIOpenCode

11 13d ago A scan tokens not measured original Apache-2.0
View the source file TheBarmaEffect/glassbox
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add mcp/thebarmaeffect/glassbox/glassbox-framework
Clone the repo
git clone --depth 1 https://github.com/TheBarmaEffect/glassbox

Made for: Claude Code, Codex, Cursor, Gemini CLI, OpenCode.

Per session not measured What this adds to a session before it is invoked.
When invoked not measured Not applicable: nothing here is loaded into a session.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Security

Grade A, and why

glassbox-framework scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

mcp/server.json · 15 lines

What it actually says

{
  "glassbox-framework": {
    "command": "npx",
    "args": [
      "-y",
      "@glassbox-framework/mcp"
    ],
    "env": {
      "ANTHROPIC_API_KEY": "",
      "GLASSBOX_MODEL": "claude-sonnet-4-6",
      "GLASSBOX_MAX_TOKENS": "2048",
      "GLASSBOX_ECS_MODE": "arithmetic"
    }
  }
}
Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 15 lines scan A 7fd24f1d476f

Subscribe to this mod's changes

glassbox-framework is an MCP server published in the GitHub repository TheBarmaEffect/glassbox (11 stars, last pushed 13d ago), licensed Apache-2.0. Its token cost is not measured: an MCP server costs its tool schemas, not its config file. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

AI SafetyBuilding agentsConstitutional AISecurityai-verificationanthropicclaudehallucination
Related

Other mcp servers, from other repositories

mettle-mcp

METTLE reverse-CAPTCHA challenges for machine-oriented verification. Runs locally from the mettle-verifier Python package.

Creed-Space/METTLE · not measured

creedspace-mcp-server

Constitutional-AI safety guardrails for any LLM — personas, constitutions, adjudication. Runs locally from the @creedspace/mcp-server npm package.

Creed-Space/creedspace-mcp-server · not measured

vaporcheck

MCP server "vaporcheck" as configured in cdmx-in/vaporcheck. Launched with py -3 c:/Users/Roney/Documents/new-project/vaporcheck/server.py.

cdmx-in/vaporcheck · not measured

verifiable-memory

Verifiable memory for AI agents — 0% hallucination, cited, provably forgettable, deterministic. MCP server. CPU-only, no GPU, zero dependencies. Runs locally from the verifiable-memory-mcp Python package.

Mars-proj/verifiable-memory · not measured

mcp

Authorize consequential AI agent actions before execution. Runs locally from the @decionis/mcp npm package. Needs 1 environment variable to run.

decionis/agent-safe-pipeline · not measured

agentgate

Human-in-the-loop approval gateway for AI agents over MCP: request, decide, approve. Runs locally from the @agentkitai/agentgate-mcp npm package.

agentkitai/agentgate · not measured