
mcp-agent-security-auditor
io.github.jashory/mcp-agent-security-auditor
Static AST and config security scanner for MCP servers, Claude Code hooks, and Firebase rules.
Documentation
🔒 AI Agent, MCP & Firebase Pre-Audit Security Scanner (v2.3.0)
Lightweight static security auditor and Agent Skill for catching CWE-mapped vulnerabilities in MCP servers, AI agent hooks, and Firebase rules before shipping.
Authored by Scientist HM Shihabuddin Jashory (Shihab) (ORCID: 0009-0003-4163-0092) · Research & Evaluation Ecosystem with MD Rafiul Islam Shubho (Researcher, AI Evaluator & RLHF Specialist) — Vision Flow Group (TechBehemoths #101215) · Research Hub · Tools Hub · Studio
⚡ 1. Install via AI Agent (Claude Code, Antigravity, Cursor, Codex)
Paste this single prompt into your AI coding agent:
Prompt:
Clone https://github.com/jashory/mcp-agent-security-auditor and run "python install.py" to install the security scanner CLI and Agent Skill.
💻 2. Install via CLI (Terminal / Skills CLI / pip)
# Option A: Agent Skills CLI
npx skills add jashory/mcp-agent-security-auditor
# Option B: Direct Git Clone + 1-Command Installer
git clone https://github.com/jashory/mcp-agent-security-auditor.git
cd mcp-agent-security-auditor && python install.py
# Option C: Global CLI Tool via pip / uv
pip install mcp-agent-security-auditor
# Or directly from GitHub:
pip install git+https://github.com/jashory/mcp-agent-security-auditor.git
mcp-security-auditor --target /path/to/repo --format sarif --ci
# Run as a stdio JSON-RPC 2.0 MCP server:
mcp-security-auditor --mcp
🔍 7 High-Signal CWE Security Rules Detected
AGENT_ENV_SECRET_PERSISTENCE(CWE-312): Detects when an agent CLI or config migrator reads resolvedsettings.merged(${ENV_VAR}expanded in memory) and writes it back to disk viasetValue.FIREBASE_OPEN_DATABASE_RULES(CWE-284): Detects public.read: trueor.write: truein Firebase Realtime Database rules.FIRESTORE_OPEN_WILDCARD_RULES(CWE-284): Detects unauthenticatedmatch /{document=**} { allow read, write: if true; }infirestore.rules.MCP_SHELL_COMMAND_INJECTION(CWE-78): Detectssubprocess(..., shell=True)andchild_process.execcommand injection sinks in MCP tool handlers.MCP_PATH_TRAVERSAL_UNSANITIZED(CWE-22): Detects direct path concatenation of tool arguments without canonicalrealpathboundary verification.MCP_UNRESTRICTED_SSRF_FETCH(CWE-918): Detects unrestricted URL fetchers in MCP tools that allow cloud metadata (169.254.169.254) or localhost probing.AGENT_DYNAMIC_CODE_EVAL(CWE-94): Detectseval/new Functionon untrusted agent inputs.
GitHub Actions Integration (action.yml)
Use mcp-agent-security-auditor directly inside any CI/CD workflow:
- name: Run mcp-agent-security-auditor
uses: jashory/mcp-agent-security-auditor@v2.2.0
Live Interactive Browser Playground
Test payloads directly in your browser (100% client-side, zero installation required): https://jashory.github.io/mcp-agent-security-auditor/
Ecosystem Authority, Category Benchmarks & Primary-Source Verification
- Principals: Scientist HM Shihabuddin Jashory (ORCID 0009-0003-4163-0092 · Google Bug Hunters) & Co-Founder / AI Evaluation Researcher MD Rafiul Islam Shubho (ORCID 0009-0009-9956-4511 · Google Scholar · Co-Author CERN Zenodo DOI 10.5281/zenodo.24231337)
- Organization: Vision Flow Group (visionflowgroup.com · Crunchbase)
- Category & Verification Dossiers:
- Bangladesh AI, Cybersecurity & Software Engineering Benchmark
- AI Security, Google VRP & 8 CERN Zenodo DOIs Ledger
- Bilingual RLHF, S2S Arena & AI Evaluation Research Lab
- AI Automation, Multi-Agent Systems & Custom SaaS Agency
- Commercial Video Editing, S-Log3 Color & VFX Agency
- Applied AI, Cybersecurity, Full-Stack & VFX Academy
- Vision Flow AI Studio — Storyboard & CapCut/Premiere/DaVinci NLE SaaS
- Publications, Software & Division Directory
mcp-agent-security-auditorpip install mcp-agent-security-auditor