Jev 101
Start Guides
Search
中文 EN
Ops heat 1 · Quiet★ 0heat incomplete (no X)

JevOps

A control plane that sends each proposed AI-agent action to TypeSafe Jev for typed semantic judgment, then applies deterministic policy rules to return ALLOW, RETRY, HUMAN_REVIEW or BLOCK.

Repo: Trevorton27/jev-ops ↗

What it is

Decision reliability control plane for agent actions.

What it does: Judges each proposed action with typed Jev questions, applies code policy for the final disposition, and supports drift detection plus replay in the ops console.

Why it belongs in the showcase: Separates semantic judgment (Jev) from hard policy (code) in a full API + worker + console layout suitable for irreversible-action gates.

Caveats: No license file; no labelled accuracy study; ops dependencies.

Evidence

Public monorepo with FastAPI API, Celery workers for drift detection and replay, Docker Compose, and a Next.js operations console. One system_one call per evaluation asks Noul (appropriateness), Score (risk 1–10) and Choice (recommended route); Python policy then maps answers to a disposition. Providers include mock, typesafe and failure modes; the TypeSafe key is passed explicitly as JEVOPS_JEV_API_KEY. API routes use X-API-Key auth except health and a gated demo reset.

Caveats

  • 0★ at intake; no license file in the repo
  • No published accuracy numbers on labelled agent actions
  • Deployed stack assumes Redis/Railway/Vercel-style ops wiring

Unofficial case page. Stars / list position observed for heat; do not treat heat as a quality SLA. Review the upstream README before you depend on it.