User repository on GitHub
- Forks
- 0
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
Jev directory
CommunityJev Tools & CLIs
ChangeGate
A confidence-aware decision gate for autonomous software changes. It runs an explicit LangGraph flow of inspect, checks and evaluate, then routes each proposal to apply, human review or stop; the decision provider can be the TypeSafe Jev API.
- Stars
- 1
- Language
- TypeScript
- Reading time
- 2 min
- Published by
- Community
- Author
- jacobjerryarackal
- Last verified
- Oct 2026
Works forJev for Decision gating
Used inJev for Software & DevTools
- Languages
- TypeScriptPython
- Platforms
- BrowserVercel AI Gateway
Highlights
- Routes each proposed change to Apply, Review Required or Stopped with a rationale.
- JEV_PROVIDER selects mock (default), typesafe (api.typesafe.ai/v1/systemone) or the jev-agent relay, keyed by TYPESAFE_API_KEY or JEV_AGENT_KEY.
- Explicit LangGraph flow: inspect, checks, evaluate, then one of three terminal nodes.
- Five built-in scenarios exercise every route, from a docstring fix to an auth bypass that must be stopped.
- The README insists it is a deterministic workflow with a bounded decision step, not an autonomous agent.
Quickstart
bash
cd backend
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --port 8003Watch out
README and GitHub metadata reviewed 2026-10-17; the digest truncates the workflow list and the repository declares no license. The five scenario proposals are predefined test fixtures, and runs hit a mock evaluator unless a provider key is configured.
Questions about this entry
Short answers drawn from the linked project or page.
A confidence-aware decision gate for autonomous software changes. It evaluates a proposed change and routes it to automatic application, human review or stop, keeping the workflow execution separate from the decision itself.
Through the JEV_PROVIDER variable: the typesafe option calls the TypeSafe AI Jev API at api.typesafe.ai/v1/systemone with TYPESAFE_API_KEY, and jev_agent uses a relay service that needs JEV_AGENT_KEY.
An explicit LangGraph flow: inspect change metadata, touched paths and change type; run automated checks such as syntax, tests and static analysis; evaluate safety and policy; then reach apply, review or stop.
The default provider is mock, a deterministic local evaluator with no external dependencies, so the five included scenario proposals run end to end without any TypeSafe or relay credentials.
The backend needs Python 3.11 or newer: create a venv, pip install -r requirements.txt, then uvicorn app.main:app --port 8003. The frontend needs Node 18 or newer: npm install and npm run dev on port 5176, proxying to the backend.
No. The README says it is not an autonomous agent or chatbot but a deterministic workflow with a bounded decision step, and execution events stream to the client in real time over server-sent events.
Five predefined proposals exercise every route: a docstring and type-annotation fix and a dependency bump apply automatically, a database column deprecation needs review, and broken tests or an unsanitized auth bypass are stopped.
More Jev resources like this
A guard that stands outside an LLM agent and uses TypeSafe Jev to score both incoming prompts and proposed actions, turning the score into a block, a question, or a pass; in the README's Gandalf test it rejected 17 of 21 published injection prompts before send.
Jev Tools & CLIs#community#github#security
Community
JamesANZ/JevPromptShieldTypeScript
JamesANZ/JevPromptShieldTypeScriptA Chrome and Edge side-panel extension that sorts your existing bookmarks so you can find a saved tool by describing its task. Import classification uses 920+ built-in site patterns with no API key; the Jev decision model is an optional re-ranker for Smart search.
Jev Tools & CLIs#community#github#chrome-extension
Community
tolson64/kitlane-extensionTypeScript
tolson64/kitlane-extensionTypeScriptA self-hosted multi-bot group chat: rooms hold people and bots, mentions wake specific bots, and with no mention Jev scores who should reply while a daemon on your own machine runs each turn through Pi.
Jev Tools & CLIs#community#github#typescript
Community
nikuscs/orbsTypeScript
nikuscs/orbsTypeScriptA per-turn model router for the Pi coding agent: a small judge picks the cheap or the strong tier for each message, with multi-model failover and task-level orchestration, plus an optional Jev judge that answers with a calibrated probability.
Jev Tools & CLIs#community#github#router
A local-first studio for building and regression-testing Jev-powered conversational state machines: define states visually, watch Jev's typed choice meet your threshold, and export runnable TypeScript.
Jev Tools & CLIs#community#typescript#state-machine
Community
priyankark/jev-stateTypeScript
priyankark/jev-stateTypeScriptA Rotten Tomatoes-style ratings site for ML papers: reader votes plus a panel of 20 AI reviewer personas that scores every paper from day one, with Jev answering 20 typed questions per paper and classifying areas and paper types.
Jev Tools & CLIs#community#github#typescript
Community
LzyFischer/good-papersTypeScript
LzyFischer/good-papersTypeScriptFrom 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.