A guard-agnostic benchmark for MCP injection / exfil / tool-poisoning detection. Versioned cases + a language-agnostic runner that scores any guard through its own published CLI.
Drift inferred · capture-to-capture
- HIGH code analysis flagged committed secret ×2 in getmcpm/mcp-guardbench
- HIGH code analysis flagged committed secret ×2 in getmcpm/mcp-guardbench
- HIGH code analysis flagged committed secret ×2 in getmcpm/mcp-guardbench
transport — counts 0 tools · 0 res
· 0 prompts
permission surface via code analysis
No tools enumerated yet for this server.
prompt-surface
shipped agent-instruction files + hidden-content / dangerous-code findings —
quoted from the analyzed source
analyzed commit a21ab71 · analyzer v33 · 3d ago
danger signals2
- committed secret GitHub token getmcpm-mcp-guardbench-a21ab71/cases/attacks/credential-egress-github-pat.json :18
ghp_A1…(40 chars, redacted) - committed secret private key getmcpm-mcp-guardbench-a21ab71/cases/attacks/credential-egress-private-key.json :18
PEM private key block (redacted)
evidence-backed
findings quoted directly from the published source artifact — not inferred
code files: 5
filesystem 3
- fs getmcpm-mcp-guardbench-a21ab71/adapters/cisco/adapter.mjs :41
import { mkdtempSync, writeFileSync, rmSync } from "node:fs"; - fs getmcpm-mcp-guardbench-a21ab71/runner/run.mjs :20
import { readFileSync, readdirSync, writeFileSync, existsSync, mkdirSync } from "node:fs"; - fs (weak) getmcpm-mcp-guardbench-a21ab71/scripts/extract-from-mcpm.mjs :17
import { readFileSync, writeFileSync, readdirSync, mkdirSync } from "node:fs";
shell / exec 3
- shell getmcpm-mcp-guardbench-a21ab71/adapters/cisco/adapter.mjs :39
import { spawn } from "node:child_process"; - shell getmcpm-mcp-guardbench-a21ab71/adapters/mcpm/adapter.mjs :27
import { spawn } from "node:child_process"; - shell getmcpm-mcp-guardbench-a21ab71/runner/run.mjs :18
import { spawn } from "node:child_process";