Agent skill · security · aradotso
agent-security-scanner-mcp
Security scanner MCP server for AI coding agents with vulnerability detection, package hallucination checks, prompt injection firewall, AST analysis, and auto-fix capabilities
Why this skill is useful
Adds executable security scanning scripts and auto-fix capabilities that fundamentally enhance the AI's ability to secure code and detect vulnerabilities.
What it needs
Requires npm installed locally. Requires anthropic, openai account access. About 7k tokens when loaded. Last updated 2026-08-04. 11 stars on the source repository.
What this skill does
agent-security-scanner-mcp Skill by ara.so — Security Skills collection. agent-security-scanner-mcp is a comprehensive security scanner designed for AI coding agents and autonomous assistants. It provides vulnerability scanning (1,700+ rules across 12 languages), package hallucination detection (4.3M+ packages), prompt injection firewall, AST + taint analysis, auto-fix capabilities, SBOM generation, and LLM-powered semantic code review. Available in two versions: ProofLayer (Lightweight): 81.5KB, 4-second install, pure regex, 400+ rules Full Version (Advanced): AST analysis, 1,700+ rules, LLM code review, Python analyzer Installation Full Version (Recommended for AI Agents) MCP Server Setup (for Claude Code, Cursor, Windsurf, Cline): Replace claude-code with your client: cursor, claude-desktop, windsurf, cline, kilo-code, opencode, or cody. Restart your client after init. ProofLayer (Lightweight) Key Tools & Commands MCP Tools (Available in AI Agents) Tool Use Case ------ ---------- scansecurity Scan code for vulnerabilities (1,700+ rules, AST + taint analysis) fixsecurity Auto-fix detected issues (120 fix templates) scangitdiff Scan only changed files in git diff scanproject Full project audit with A-F security grading checkpackage Verify package isn't AI-hallucinated (4.3M+ packages) scanpackages Bulk-check all imports in a file scanagentprompt Detect prompt injection (59 rules + multi-encoding bypass detection) scanagentaction Pre-execution safety check for agent actions (ALLOW/WARN/BLOCK) scanmcpserver Audit MCP server source for security issues scanskill Deep scan of OpenClaw skills (ClawHavoc malware signatures) sbomgenerate Generate CycloneDX v1.5 SBOM (8 lock file formats) sbomscanvulnerabilities Cross-reference SBOM against OSV.dev for CVEs evaluatecompliance Evaluate project against SOC2/GDPR frameworks CLI Commands Configuration Environment Variables MCP Server Configuration After running npx agent-security-scanner-mcp init <client>, configuration is auto …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills aradotso/agent-security-scanner-mcp