Show HN: I scanned 87 MCP servers for agent-authority hygiene – leaderboard

Category: security

Tags: ai-security, capability-based-security, mcp

Score: 8.5/10 (Innovation: 8, Technical: 9, Documentation: 9, Utility: 8)

Capframe is a deterministic security platform for AI agents that interact with tools, addressing prompt injection and confused deputy attacks through capability tokens and runtime policy enforcement. It innovatively combines tool discovery, scoped authorization, and guard enforcement into a unified CLI, with a unique self-audit culture and a custom risk taxonomy (CAST). The project fills a critical gap for production-grade agent security.

Target audience: backend devs, devops, security engineers

Repository: https://capframe.ai/leaderboard · Rust · MIT

View on Hacker News