发现优质的 AI Agent 技能
聚合 Claude Skills、LangChain、AutoGPT 等优质资源，助力开发者快速构建智能应用
Model Switchboard v2.0 — Security & Code Audit
**Auditor:** Claude Opus 4.6 (subagent) **Date:** 2026-02-25 **Scope:** All source files (switchboard.sh, validate.py, redundancy.py, model-registry.json, task-routing.json, ui/index.html, SKILL.md, README.md) **Verdict:** **CONDITIONAL PASS** — No critical vulnerabilities found. Several medium/low issues to address before ClawHub publish.
Security Considerations
This skill interacts with Twitter/X using authenticated sessions. Please read and understand these security implications.
MoltThreats
MoltThreats is the agent-native security signal layer by PromptIntel.
Security Audit Report - AdGuard Home Skill
2026-02-24 (Updated: 2026-02-24)
Security Audit Report - AdGuard Home Skill
2026-02-24 (Updated: 2026-02-24)
Cyber Security Engineer
Implement these controls in every security-sensitive task:
skill-guard
**The only pre-install security gate for ClawHub skills.**
Skill Vetter 🔒
Security-first vetting protocol for AI agent skills. **Never install a skill without vetting it first.**