Show HN: Stop your AI agents from approving their own work
Category: security
Tags: ai-agents, governance, security, audit, role-based-access-control, human-in-the-loop, typescript
Score: 8.3/10 (Innovation: 8, Technical: 9, Documentation: 9, Utility: 7)
MakerChecker is a self-hosted governance layer for AI agents that enforces role-based access control, limits, and human approval gates before any tool call executes, with a tamper-evident Ed25519-signed audit chain. It combines structural enforcement at machine speed with cryptographic auditability, addressing a critical gap in AI agent safety and compliance.
Target audience: backend devs, security engineers, infrastructure engineers, data engineers
Repository: https://github.com/sammysltd/makerchecker · TypeScript · AGPL-3.0
View on Hacker News