What happens when Jev is unavailable or low-confidence?

The plugin delegates to stock dsh behavior by calling next(). A missing key, provider cooldown, timeout, parse mismatch or low-confidence answer never guesses and never widens permissions, so a dead provider costs no behavior.

How do I roll the plugin out safely?

Install it and leave mode: shadow, which records every decision with full probability distributions and enforces nothing. Watch /jev-stats, compare the counterfactuals, then switch to mode: enforce. Commands persist to the profile on the next tool call.

What does one gate request cost and decide?

One request covers risk class, irreversibility, task-match and injection suspicion, at roughly $0.00002 measured over 501 input tokens. Confident high-risk calls are denied, ambiguous ones escalate to a human, and clearly granted reversible ones stop asking for approval.

What do the published sample numbers look like?

The README's /jev-stats sample shows 41 calls with 0 degraded and 9 cached, actions of delegate 33, ask 6 and deny 2, 18,234 input tokens for an estimated $0.000766, and latency of p50 247 ms, p95 512 ms and max 611 ms.

How does the session-retention feature work?

The plugin takes over the upstream session-reference row and scores every droppable message as noise, background, relevant or critical in one Jev fan-out, making the drop order least valuable first with truncation cutting padding before substance. Checkpoints and the newest message stay protected and the byte budget is honored exactly as upstream, and with sessionReferenceEnabled false the row renders byte-for-byte like stock.

How does it treat suspected prompt injection?

Tool arguments enter the Jev state data field only, with instructions fixed as strings, and a suspected-injection answer escalates rather than suppressing. The README notes Jev's maker acknowledges adversarial inputs can sway classifiers, so denial here is treated as an accelerator and never the last line of defense.

What license and runtime does it need?

MIT-licensed, published to npm as dsh-jev-interceptor, and it needs Node 20.3 or newer plus a TypeSafe or OpenRouter key. The 77 tests run with no network, and every decision lands in a telemetry.jsonl under the harness home, honoring $DSH_HOME.

CommunityJev Tools & CLIs

Dsh Jev Interceptor

A DeepSeek Harness plugin where Jev Choice and Noul checks gate every non-read-only tool call - deny confident high-risk calls, escalate ambiguous ones, auto-approve clearly granted reversible ones - and optionally score session messages so valuable context survives instead of being dropped oldest-first.

See highlights

Open github.com

Stars: 27
Language: TypeScript
License: MIT
Published by: Community
Last verified: Oct 2026

AskTheWay/dsh-jev-interceptor

User repository on GitHub

View on GitHub

⚡ Millisecond System-1 judgement for every tool call in DeepSeek Harness — Jev-powered risk classification & evidence-gated auto-approval. Fail-closed by construction. dsh 生态第一个 System-1 决策插件

Forks: 3
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

What it does, in one breath: before a tool call runs, Jev classifies its risk, irreversibility, task-fit, and injection-suspicion in one $0.00002 request — confident high-risk calls get denied, medium ones escalate to a human, and clearly-granted reversible ones stop wasting your clicks on approval dialogs. And when an @session snapshot gets injected, Jev scores every message's value for the citing task so the error traceback survives the byte budget instead of the oldest small talk. Every doubt, every timeout,…

Quoted from the project's README. Read the full README.

Tagscommunity plugin guardrails deepseek

Works forJev for Decision gating Jev for Tool calling Jev for Verification

Used inJev for Software & DevTools Jev for Security & Privacy

Languages: TypeScript
Capabilities: Open sourceRequires API key

Highlights

One request covers risk class, irreversibility, task-match and injection suspicion for about $0.00002 (measured 501 input tokens).
README sample /jev-stats output: 41 calls, p50 247 ms, p95 512 ms, actions delegate=33 ask=6 deny=2.
Starts in shadow mode recording every decision with probabilities and enforcing nothing; flip to enforce after reviewing stats.
Never allows: 'no objection' is expressed as next(), so downstream listeners keep their veto, and failures degrade to stock dsh behavior.
Semantic session retention replaces oldest-first dropping with least-valuable-first scoring; 77 tests run with no network.

Watch out

MIT-licensed and fail-closed, but it tracks the DeepSeek Harness plugin API and needs a TypeSafe or OpenRouter key. Nothing is auto-approved until tools are listed in the allowlist.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

X: Agent Handoff Gate: agents verify what they hand off

X: Jev as an LLM guardrail inside Agentgateway

X: Vercel's fx safety reviewer, measured at 18× faster

Reddit: Early experiments wiring Jev into Pi as a safety layer and model router.

Related terms

Glossary definitions related to this entry.

Guardrail

More Jev resources like this

DeepSeek Harness Jev Plugin

203

Native DeepSeek Harness (DSH) plugin that uses Jev or Luna Decisions as a System One judgment layer across 12 separately configurable features: skill selection, file ranking, drift reminders, completion checks, goal supervision, corrections, log admission and workspace approvals. All features ship disabled by default.

Jev Tools & CLIs#community#deepseek#plugin

Community

luobosibing2/dsh-jev-pluginJavaScript

pi-verdict

Permission gate extension for the pi coding agent: deterministic allow/deny rules and protected paths settle clear cases first, then gray-zone tool calls get one structured classify() verdict; Jev is an optional native classifier via TypeSafe direct, OpenRouter, or other gateways, failing closed on error.

Jev Tools & CLIs#community#guardrails#permissions

Community

jesset/pi-verdictTypeScript

Edward

Supervisor for unattended coding agents: it tracks cross-turn state and asks one batched Jev Choice whether a trajectory should continue, pause or escalate, while deterministic code owns dangerous-command blocking, budget caps and Ed25519-signed receipts.

Jev Tools & CLIs#community#guardrails#safety

Community

VeridicalTech/EdwardPython

openclaw-jev-leakguard

OpenClaw plugin that checks every outgoing agent message against its destination channel tier, running local key-format and term rules and then five Jev Noul questions in one call through OpenClaw's decisionModel, which can point at hosted Jev or a local Kev.

Jev Tools & CLIs#community#openclaw#guardrails

Community

yousan/openclaw-jev-leakguardJavaScript

approval-judge-bridge

OpenAI-compatible /v1/chat/completions proxy that answers an agent host's approval-guardian call with APPROVE, DENY or ESCALATE. Jev answers one Choice question with a calibrated distribution; missing keys, timeouts and malformed answers all fail closed to ESCALATE.

Jev Tools & CLIs#community#agent-safety#guardrails

Community

oppih/approval-judge-bridgePython

agent-chaperone

An MCP proxy and hooks adapter that screens an agent's tool calls before they run and the tool results before the agent reads them, with thresholds in a policy file.

Jev Agents & MCP#community#mcp#security

Community

agent-chaperone/agent-chaperoneTypeScript

From the Jev guides

Original write-ups on what this entry is for.

Jev on the Gateways: Vercel, Workers AI, and OpenRouter

Oct 12, 2026

The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.

Jev Tools & CLIs#gateway#vercel#cloudflare

Read article

Exporting Decision Models: GGUF, ONNX, and Device Runtimes

Oct 10, 2026

Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.

Jev Tools & CLIs#open-models#local#onnx

Read article

Records at Rest: Documents, Inboxes, and Risk Queues

Sep 28, 2026

Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.

Jev Tools & CLIs#classification#email#security

Read article

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai