Show HN: Govern Anthropic Managed Agents with 3 lines of code

Category: ai-ml

Tags: ai-governance, agent-security, audit-trail

Score: 7.0/10 (Innovation: 7, Technical: 6, Documentation: 8, Utility: 7)

AgentLattice provides governance and audit capabilities for Anthropic's Managed Agents, intercepting tool calls (bash, file writes, web fetches) for policy evaluation and human approval before execution. It's interesting because it addresses a critical security and compliance gap in autonomous AI agents with a simple API integration and tamper-proof audit trails.

Target audience: backend devs, devops, ai engineers

Repository: https://github.com/AgentLattice/govern-anthropic-demo · TypeScript · MIT

View on Hacker News