Skill Control Plane - Governance for AI agent skills — MCP-native

Hey Product Hunt 👋 — Benjamin here, solo maker. I built Skill Control Plane because AI "skills" (the Claude Skills / Agent Skills / MCP-tool plugins that agents use to do anything useful) are exploding, and nobody is governing them. Think of skills as the new npm. Same attack vector. Same supply-chain risk. Except nobody at your company has visibility into what skills your AI agents are installing, what they do, or whether they've been scanned. Skill Control Plane is the missing layer: 🧬 IMPORT — from GitHub, filesystem, or Claude's built-in library 🔍 SCAN — 3 automated layers (pattern detection, credential entropy, AI deep-scan via Claude on AWS Bedrock) ✅ APPROVE — configurable risk thresholds, RBAC (4 roles, 18 permissions), audit trail 🚀 DISTRIBUTE — MCP server, one-click export to Claude Code / Cursor / Windsurf / any MCP agent 📊 TRACK — usage analytics, human-vs-agent differentiation, immutable log Pricing: Free for 25 skills + pattern scan (real free tier, no 14-day countdown). Pro is $299/mo for teams with full scanning + approval workflows. Enterprise has SSO/SAML/on-prem. I built this solo over the last 6 months because I'm convinced "Agent Skill Governance" becomes its own category in 2026 — same way "Container Security" became a category after Docker took off. I'll be online all day answering questions. Try the free tier, break it, tell me what's missing. Roast the security model, grill the approval UX, tell me the tagline is bad. Upvotes help me reach more builders — thank you for even reading this far. 🙏 → skillcontrolplane.com → 2-min demo: https://youtu.be/9qoxh6ox-z4?si=...