タグ：claude-code-hooks

Claude Code / Codex / pi plugin that hands agent steps needing no text output to Jev (TypeSafe's judgment model) — measured p50 ~230 ms and ~$0.02 per 1,000 judgments, with typed escalation back to the LLM

A Claude Code plugin that keeps your AI agents honest. Claude plans tasks as phases with checks, and when it tries to stop, Jevflow runs them and asks Jev if the work is really done. Many agents, one plan, with a live viewer to watch it all.

Cut Claude Code's skill manifest by ~75% with TypeSafe Jev. Scores every installed skill for relevance and hides the rest via skillOverrides — 12,750 → 3,185 tokens on a 217-skill install, for $0.0009 a session.

Pre-execution risk gate for AI coding agents (Claude Code, Codex, opencode, pi, Hermes). Blocks or asks on risky shell commands, guards against prompt injection in what agents read, escalates to System 2 and a human queue. Optional TypeSafe Jev.

A Claude Code hook that asks whether the decision you are writing needs a model at all. Includes a measured 149-row comparison of TypeSafe Jev against Claude Haiku 4.5.

Event-driven trajectory watchdog for coding agents: per-event LLM judgments (Jev, Claude, GPT) accumulated into quarantine decisions