Show HN: I built a safety gate for autonomous coding agents

Category: security

Tags: ai-safety, prompt-injection, coding-agents, security, hooks

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 7, Utility: 7)

Inbin Gate is a safety hook for coding agents (IBM Bob, Claude Code) that intercepts terminal commands and protected file writes, allowing only actions explicitly authorized by the developer (signed intent) or repository maintainers. It derives authority from the provenance channel rather than the content, providing structural resistance to prompt injection and unsafe autonomous actions.

Target audience: backend devs, devops, security engineers

Repository: https://inbin-gate.vercel.app/ · JavaScript · MIT · 1 stars

View on Hacker News