Article
CommunityJev Guides & Articles
A Jev gate belongs in a hook, not MCP
Automater's runbook for turning a Claude Code plus Jev check into enforcement: sort integrations by who invokes them, put the gate in a PreToolUse hook that fails closed with deny rules behind, and keep MCP for lookups.
- Format
- Article
- Reading time
- 3 min
- Published by
- Community
- Author
- Russell Winslow
- Last verified
- Oct 2026
- Platforms
- MCP
Highlights
- Core claim: an MCP tool the model can decline to call is a second opinion, whatever its name says.
- Prescription: one PreToolUse hook contract that denies on a risky verdict and on every Jev failure, with deny rules behind it.
- Recounts launch week: the official typesafe@typesafe-ai skill plugin, plus MCP servers jkudish/jev-mcp (227 stars) and itsmostafa/typesafe-mcp (about 200).
- Also names LangChain's AutoModeMiddleware blocking tool calls at or above 0.5 and Vercel's eve sending a Jev caution verdict to a human.
- Notes that a hook fails open by default when it crashes or stalls, so a permission-and-sandbox wall backs it up.
Quickstart
bash
claude plugin marketplace add typesafe-ai/skills
claude plugin install typesafe@typesafe-aiWatch out
Grounded on the captured article text, dated September 21, 2026 and credited to Russell Winslow, plus the digest's meta description; the capture is truncated inside Step 1's inventory table, so the hook contract, the Codex version and the permission-and-sandbox section are described only as the article promises them. Star counts are as of September 21, 2026 and will have moved.
Questions about this entry
Short answers drawn from the linked project or page.
That a Jev MCP tool runs only when the model chooses to call it, so it does not run when the model is confused, rushed, injected or sure of itself, the four moments you wanted a gate for. The runbook's answer is to put the gate in a PreToolUse hook that fails closed, keeping MCP for lookups.
TypeSafe launched Jev on September 15, 2026, and by September 21 the official integration was a skill plugin, typesafe@typesafe-ai, about 1.5K stars, that teaches question types and makes no tool calls. The MCP route was community-built: jkudish/jev-mcp at 227 stars, itsmostafa/typesafe-mcp around 200, plus jev-bridge and at least nine others.
Two commands, per the article: claude plugin marketplace add typesafe-ai/skills and claude plugin install typesafe@typesafe-ai, then invoke it with /typesafe:typesafe-ai or by asking for the skill in a prompt. The article notes the page mentions no MCP server.
Because what separates a lookup from a gate is who decides the check runs, and with an MCP tool that is the model. A check the actor can decline to invoke is a suggestion with a latency budget, whatever the tool is named, even one called jev_gate.
LangChain's experimental AutoModeMiddleware, alpha September 17, asks a Jev Noul whether each listed tool call is risky and blocks at or above 0.5 while stating it does not request human approval. Vercel's eve sends a Jev caution verdict to a human, and a practitioner's open-source Auto-Guard runs Jev on every coding-agent tool call.
A hook fails open by default when it crashes or stalls. The runbook's answer is one PreToolUse hook contract that denies on a risky verdict and on every Jev failure, plus a permission-and-sandbox wall for the day the hook fails open, and a Codex version of the same contract.
Russell Winslow, published on Automater Intel on September 21, 2026, at 19 minutes of reading time. Star counts and repo details in this entry are as of that capture date and will have moved since.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: Jev in front of the HeyGen MCP
X: Playwright CLI + Jev vs Playwright MCP
X: A local MCP server that cuts computer use from 23s to 0.8s. Cartwright built Cortex, a local MCP server where…
X: A side panel that drives any site's WebMCP tools
More Jev resources like this
A portable Agent Skill that teaches coding agents - Claude Code, OpenAI Codex, Cursor and other compatible hosts - to use Jev as a bounded judgment layer, plus an optional MCP bridge and workflows for code, browser research, Blender and Unreal.
Jev Agents & MCP#community#catalog-import#logicrw
Documentation hub for wiring Jev into Hermes Agent, Nous Research's autonomous agent framework: Choice, Noul and Score arrive as MCP-style tools, with context curation, turn admission and receipt-backed verification described end to end.
Jev Agents & MCP#community#catalog-import#madewithjev
Community
andyholst/hermes-typesafe-jev
andyholst/hermes-typesafe-jevOllaya's documentation index for its local decision-model runtime: quickstart, the full CLI reference from run to mcp, native and TypeSafe-compatible API endpoints, Modelfiles, and MCP plus agent-skill guidance.
Jev Guides & ArticlesDocs#community#catalog-import#cobanov
Community
Jev-backed web search for Claude Code and Codex in three forms: a function-hook plugin that answers the built-in WebSearch with Jev, an MCP server, and a CLI. Jev picks sources and time window, then ranks results by relevance; zero runtime dependencies.
Jev Agents & MCP#community#catalog-import#madewithjev
Community
mukiwu/jev-search-mcpJavaScript
mukiwu/jev-search-mcpJavaScriptMCP server that lets Claude Code, Codex and other agents drive a real Chrome browser while JEV makes the small typed decisions — which element, which suggestion, did the click work — with calibrated confidence driving act, verify or ask.
Jev Agents & MCP#community#catalog-import#logicrw
Community
legostin/jev-mcpTypeScript
legostin/jev-mcpTypeScriptThe Go server, agent CLI and MCP for Laya - open horizontal System 1 decision models that score structured features and return ranked pattern probabilities instead of text. It ships a JEV compatibility API and a migration guide.
Jev Repos & SDKs#community#catalog-import#madewithjev
Community
neko233-com/laya-goGo
neko233-com/laya-goGoFrom 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.