User repository on GitHub
- Forks
- 0
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
Jev directory
CommunityJev Integrations & Adapters
Jev for Amp
A small advisory plugin for Amp that compares supplied alternatives against supplied evidence using Jev. It returns typed choices and probabilities — not research, explanations or permission to act; the author labels it a V1 pilot, not an Oracle replacement or a correctness gate.
- Stars
- 1
- Language
- TypeScript
- Reading time
- 3 min
- Published by
- Community
- Author
- thesammykins
- Last verified
- Oct 2026
- Languages
- TypeScript
- Capabilities
- Open sourceRequires API key
Highlights
- Registers one tool, jev_evaluate, and is labelled a V1 pilot rather than an Oracle replacement or a correctness gate.
- Only index.ts and decision.ts are needed at runtime; there are no runtime package dependencies.
- Reads the JEV_KEY secret Amp injects, falling back to TYPESAFE_API_KEY; JEV_KEY takes precedence when both are set.
- Blocks its configured token when found verbatim in input, but the README warns this is not a general secret scanner.
- Checked on Linux/x64 with Bun 1.3.10; after adding an orb secret, run amp orb restart-processes.
Quickstart
bash
test -n "$JEV_KEY" && echo 'JEV_KEY available' || echo 'JEV_KEY missing'Watch out
Copy reviewed against the repo README on 2026-10-15; no benchmarks are published and the GitHub metadata lists no license. Development was verified only on Linux/x64 with Bun 1.3.10 and pinned Amp API types, and the README stresses the advisory is guidance, not a technical quota or an automatic usage gate.
Questions about this entry
Short answers drawn from the linked project or page.
It is a small advisory tool for comparing supplied alternatives against supplied evidence. Jev returns typed choices and probabilities; it does not do research, provide explanations or grant permission to act, and the author labels it a V1 pilot rather than a correctness gate.
A single tool, jev_evaluate. The README's install prompt asks you to confirm that jev_evaluate is the only tool the plugin registers, and only index.ts and decision.ts are needed at runtime — there are no runtime package dependencies.
Into Amp's native secrets UI as JEV_KEY, scoped to the target orb or project; the plugin reads the environment variable Amp injects. In an orb, run amp orb restart-processes after adding or changing the secret, then reload the plugin if needed.
The README says not to use it for routine choices, reassurance, open-ended research, generating missing designs, guessing what a user will authorize, or proving a system safe — inspect sources, run tests, ask the user or use a reasoning model for those.
It never takes a key as a tool argument, and no plaintext jev.apiKey setting is supported. It does block its configured token when found verbatim in input, but the README warns this is not a general secret scanner, so sanitize your own project context.
Amp with directory-plugin support and its Bun plugin runtime. Development was checked on Linux/x64 with Bun 1.3.10 and the pinned Amp API types in package.json; do not install the whole checkout as loose plugin files — only index.ts and decision.ts belong in .amp/plugins/jev/.
One call per unchanged decision; the README says not to repeat calls to obtain endorsement. That is agent guidance, not a technical quota or an automatic usage gate, so nothing in the plugin enforces a call limit.
More Jev resources like this
An OpenCode plugin that keeps the session on a cheap sticky parent model so the cache stays warm, and asks Jev whether a turn is hard before escalating through jev_escalate to temporary stronger subagents.
Jev Integrations & Adapters#community#catalog-import#cobanov
Local tool-routing classifier for coding agents: a gateway and MCP tool that ask Jev which tool to call next, keep your agent's own model and login, and log every decision so you can start by observing.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
felpsdev/jev-classifierTypeScript
felpsdev/jev-classifierTypeScriptA fork of codex-with-chatgpt that adds TypeSafe Jev plus deterministic rules to decide when Codex hands planning, debugging or review to web ChatGPT and when it switches back. Smart switching is off by default, and with it off the tool behaves exactly like the original.
Jev Integrations & Adapters#community#catalog-import#logicrw
An OpenClaw plugin that asks Jev two questions before the language model runs: whether a group message deserves a reply, and which of three configured model tiers should answer it. The reply gate is on by default; the model router is off.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
herval/openclaw-jev-pluginTypeScript
herval/openclaw-jev-pluginTypeScriptA provider-agnostic decision gateway that puts Jev in front of expensive generative LLMs: small bounded questions are batched through Jev, a deterministic policy layer builds the execution plan, and generation runs only when the policy says so.
Jev Integrations & Adapters#community#catalog-import#logicrw
A host-agnostic routing layer that sits between a coding harness and its models, choosing capacity per round from trajectory evidence like tool calls, failures and context pressure, running as a user-scoped daemon and local OpenAI-compatible proxy.
Jev Integrations & Adapters#community#catalog-import#logicrw
Community
vizuh/sabiTypeScript
vizuh/sabiTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.