Show HN: Behavioral gauge for Claude Code sessions

Category: observability

Tags: ai-agents, observability, claude-code, developer-tools, typescript

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

ccverdict is a local behavioral gauge for Claude Code sessions that monitors agent activity and provides real-time health verdicts like 'progressing' or 'drifting'. It's interesting because it fills a niche gap in AI agent observability by detecting retry loops, unchecked edits, and context pressure without relying on cloud infrastructure.

Target audience: backend devs, devops, AI/ML engineers

Repository: https://github.com/softcane/ccverdict · TypeScript · MIT · 4 stars

View on Hacker News