jevbooks

← All projects

rh-guard

24601/rh-guard · Homepage

Reward-hack radar for coding agents: structural denies plus Jev sidecar.

Verification & Guardrails95%Runner-up: Evaluation & Benchmarking

What it is

rh-guard sits in coding-agent hooks (Claude Code, Cursor, Codex, Grok Build, Pi, Amp, Prime Agent, DeepSeek Harness) and blocks reward-hacking tool use—tampering with graders, hidden tests, or the eval process—while steering toward checks the agent cannot game.

How it uses Jev

The README describes a TypeSafe Jev System One sidecar alongside structural denies, but does not specify which decisions Jev makes, on what state, or how the result is used in code.

View on GitHub

judged by Jevjev-1.13.0

Evidence

Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.

  • Jev-centricno0.21
  • Shows a System One patternno0.21
  • Handles uncertaintyno0.05
  • Measuredno0.05
  • Runnableyes0.93
  • Worth recommendingno0.19
  • Model replicano0.10
  • Problem scopescore on a 0–2 scale1.38
  • About Jevyes0.90

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 21 Sept 2026.