The request asks what to use for scanning an AI component (MCP server, agent skill, or npm/PyPI package) for security issues before trusting it. SkillTotal is the product that does exactly this: a free static scanner for MCP servers, agent skills and packages, with an MCP server door 1. Alternatives that also scan MCP servers or agent components include Ship Safe, ProofLayer and GuardVibe 346.
- 1
SkillTotalskilltotal.aiwebsite read, tools not yet checked
Free static scanner for MCP servers, agent skills and npm/PyPI packages that finds prompt injection, tool poisoning and exfiltration with file:line evidence, available as an MCP server 1.
official:MCP server · SkillTotal+1 community tools built for it
- 2
Ship Safeshipsafe.shverified and signed
Free CLI with 29 AI security agents that scan codebases, including MCP server manifests and configuration, for LLM vulnerabilities and secrets 3.
+2 community tools built for it
for agents: scan_repo · ship_safe_scan_mcp · scan_secrets
- 3
- 4
GuardVibeguardvibe.devwebsite read, tools not yet checked
MCP-native security scanner with 624 rules and 39 tools that checks AI-assisted code and works with Claude Code, Cursor and Gemini CLI 6.
Sources
- 1MCP Server & Agent Skill Security Scanner · SkillTotalskilltotal.ai · read 2026-10-29
- 3Ship Safe — AI Agent Security Scanner for Developersshipsafe.sh · read 2026-10-29
- 4AI Red Teaming & MCP Security | ProofLayerproof-layer.com · read 2026-10-29
- 6GuardVibe — MCP-Native Security Scanner for AI-Assisted Developmentguardvibe.dev · read 2026-10-29
For agents: the 10 actions behind this answer
· SkillTotal website read, tools not yet checked
Free static scanner for MCP servers, agent skills and npm/PyPI packages. Finds prompt injection, tool poisoning and exfiltration with file:line evidence.
snyk/actions · Snyk verified and signed
Check your applications for vulnerabilities using Snyk
scan_repo · ship-safe verified and signed
Run a full multi-agent security scan on a repository or directory. Runs all 20+ ship-safe security agents (injection, auth bypass, secrets, supply chain, LLM security, etc.) and returns a structured findings report with severity ratings and remediation advice. Use this when the user asks to audit, s
· ProofLayer website read, tools not yet checked
Continuously red-team LLMs and AI agents, scan MCP servers, and protect MCP traffic at runtime. Turn verified findings into audit-ready security evidence.
semgrep · semgrep verified and signed
Semgrep catches security vulnerabilities in real-time and guides Claude to write secure code from the start.
· GuardVibe website read, tools not yet checked
Security infrastructure your AI can't be — deterministic, post-cutoff-current, whole-repo, author-independent. Security MCP for vibe coding. 624 rules, 39 tools, zero config. Works with Claude Code, Cursor, Gemini CLI, and more.
scan_mcp_server · Compuute MCP Security Scanner verified and signed
Scan a public GitHub MCP-server repository for security issues. Clones the repo (shallow, <60s, <200 MB), runs compuute-scan v0.6.2 in static analysis mode (no code execution from the target), and returns a structured report with severity counts, a 0-100 score, and the 10 most sever
VulnCanon Source Scan · Viridis MCP verified and signed
Match source code against the public VulnCanon catalog of AI-agent vulnerability classes (SSRF in fetch tools, prompt injection patterns, tool-poisoning, cross-agent state pollution, etc.). Returns matched entry IDs, severity, occurrence line, and mitigation references. Backed by T-IB-05 Canon Compr
scan · safety verified and signed
semgrep · semgrep verified and signed
Runs a Semgrep security scan over a codebase: detects languages, selects rulesets, presents the plan for explicit approval, then runs every approved ruleset through scripts/run-scans.sh, which batches the semgrep processes and writes scans.json, and merges the output to SARIF. Supports two scan mode
Labels are the instruments' verdict and are never changed by ranking. Agents and robots get this same answer as JSON at /api/v1/find?q=…&answer=1 and through the MCP server.