Jevify MCP Server — AI skill for Claude Code
Local, offline MCP server for jevify: find the AI calls that are really decisions, then plan their migration to JEV in shadow mode.
How to install Jevify MCP Server
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open lucioamor/jevify-mcp-server and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Jevify MCP Server does
Local, offline MCP server for jevify: find the AI calls that are really decisions, then plan their migration to JEV in shadow mode.
Alternatives in AI
- Flow Next — Plan-first AI workflow plugin for Claude Code, OpenAI Codex, and Factory Droid 569 ★
- ZotSeek — AI semantic search for Zotero, with a built-in MCP server for AI agents (Claude Code, Codex) 193 ★
- Squeez — Hook-based token compressor for 5 AI CLI hosts (Claude Code, Copilot CLI, OpenCode, Gemini CLI, Codex CLI) 188 ★
README
jevify MCP server
Local, offline [MCP](https://modelcontextprotocol.io) server for **jevify**: it finds the AI calls in your code that are really bounded decisions (a queue, a score, a yes/no) and plans their move to [JEV](https://docs.typesafe.ai) in shadow mode.
- Deterministic. Static analysis only, with no model calls. The same input always gives the same report.
- Offline. It reads only what your agent sends it. No network, no account, no storage.
- Read-only. It never edits code. The
migratetool returns a plan, and your agent applies it only after you approve it.
It is the engine behind the [jevify skills](https://github.com/lucioamor/jevify) and the hosted service at `jevify.lovable.app`, packaged to run on your machine.
Tools
| Tool | What it does |
|---|---|
audit_files |
Audits up to 80 supplied files for runtime AI call-sites and returns a report: each finding at path#line, its classification (Generation, Choice, Score, Noul, deterministic code, embedding search, human review), confidence, risk and coverage. |
classify_ai_callsite |
Classifies one snippet and explains the evidence, the missing context and the next step. |
migrate |
Plans the move of one JEV candidate to a Choice, Score or Noul decision: native request, TypeScript composition behind an off | shadow | on flag that defaults to shadow, thresholds, fallback to the current AI path, boundary cases, validation and rollback. |
Install
Requires Node.js 20 or newer. Published on npm as [`@nxlv-ai/jevify`](https://www.npmjs.com/package/@nxlv-ai/jevify); `npx` fetches and runs it, so there is nothing to install by hand.
**Claude Code**
claude mcp add jevify -- npx -y @nxlv-ai/jevify
**Codex**
codex mcp add jevify -- npx -y @nxlv-ai/jevify
**Cursor, Claude Desktop and other clients** (`mcpServers` config)
{
"mcpServers": {
"jevify": {
"command": "npx",
"args": ["-y", "@nxlv-ai/jevify"]
}
}
}
Related Skills
Stepwarden
Every tool call your agent makes, checked before it runs. A Claude Code plugin that uses TypeSafe AI's Jev to
Plan Mode Arken
Create detailed technical implementation plans for Arken AI development. Ask clarifying questions one at a tim
Reflex Cu
Computer-use MCP server for coding agents: local macOS, local Windows, or remote Windows over SSH, with a fast
Jev MCP
TypeSafe AI Jev MCP server for bounded probabilistic decisions: four read-only stdio tools for choices, routin
AI Surface
Find and govern AI attack surfaces in application code, at PR time and inside your AI coding tool (MCP server,
Agent Toolbox
Download every AI skill, agent, plugin and MCP server you find without breaking your coding agent. Tools stay
Related Agents
Storage Manager
Use for local storage implementation - settings persistence, IndexedDB for large data, offline support, and da
Backlog MCP Validator
Validate the backlog FastMCP server against the CLI. Calls MCP tools natively via the agent-scoped backlog ser
Agy Runner
Drives one delegated task through the agy MCP server and reports back, so the orchestrator is not blocked whil