Jev directory
CommunityJev Agents & MCP
Jevify — Kevin-Zhouu
Claude Code and Codex skill that audits a codebase for LLM call sites, proposes which routing, classification, and rubric-scoring decisions to convert to Jev, and reports latency and cost comparisons after the scope is approved.
- Stars
- 4
- Reading time
- 2 min
- Published by
- Community
- Added
- Oct 2026
This is one of several Jev listings for the same project. The canonical entry is Jevify (investigation prompt).
Highlights
- Installs with npx skills add Kevin-Zhouu/jevify, then runs as /jevify in Claude Code or $jevify in Codex.
- Audits a codebase for LLM call sites and proposes routing, classification, and rubric-scoring decisions to convert.
- Scope options include a new branch, a separate copy, report only, or converting individual calls.
- Keeps free-form generation with your LLM and adds a fallback for uncertain decisions.
- Demo run: 12 questions, 136 lines of JSON; Jev 0.44s versus Claude Haiku 4.5 at 5.11s in one synthetic example.
Questions about this entry
Short answers drawn from the linked project or page.
Install with `npx skills add Kevin-Zhouu/jevify`. Then in Claude Code (beta) use `/jevify`, in Codex use `$jevify`, or say "jevify this repo." You can scope it to a path, for example `/jevify src/agents/triage.py`.
No. No key or setup questions are needed to see the plan. Live validation uses `TYPESAFE_API_KEY` or `OPENROUTER_API_KEY`; without a Jev key, the audit and approved code changes still work but live validation is skipped.
Decisions such as routing, classification, and rubric scoring are candidates for conversion. Free-form generation stays with your LLM, and you receive a conversion plan plus a results report.
The flow is audit first, then you approve scope and output before any conversion. You choose a new branch, a separate copy, report only, or individual calls. It keeps generation with your LLM, adds fallback for uncertain decisions, and runs tests.
Claude Code support is beta with known conversion validation failures, and full cross-harness re-evaluation is pending. Historical Codex DEV runs passed. The audit-first flow is described as new.
Skill helpers use Python's standard library. Development checks require Python 3.10+, Node 18+, and `npm ci` in the `eval/` directory, followed by the unittest discovery and freeze verification commands.
No. It is one synthetic example via OpenRouter, replayed from real stream timestamps: 12 questions, 136 lines of JSON, Jev at 0.44s and Claude Haiku 4.5 at 5.11s, with all 12 category choices matching. It is not a general benchmark.
Reactions & coverage
Posts, threads, and videos about this entry from around the web.
X: A usage router for Grok Bot
X: A prompt box that fills itself in
X: How Jev makes agents faster and cheaper
YouTube: CJ from Syntax demos browser use, classification, code review, a model router, and a chat bot with no LLM.
More Jev resources like this
Claude Code plugin that puts Jev behind six hooks: request triage, automatic effort, a destructive-command guard, rule review, and a final completion check. One file, zero dependencies, MIT licensed.
Jev Agents & MCP#claude-code#guardrails#coding-agent
CommunityCivalski/claujev
Claude Code and Codex skills plus a study course for building with Jev, TypeSafe's System One model. The materials pull from Jev's canonical cookbooks and docs, and the Ratings folder lists rated Jev projects and an evaluation methodology.
Jev Agents & MCP#community#github#skills
Community
tiffygk/jev-modeHTML
tiffygk/jev-modeHTMLA skill for Codex agents that splits development work into Bounded Tasks and produces a Dispatch Plan recommending a model, reasoning effort and tools per task, with optional Jev-powered assessment and an Agent fallback when Jev is unavailable or low-confidence. MIT-licensed.
Jev Agents & MCP#community#github#javascript
Community Codex skill that keeps the official TypeSafe System One guidance and adds a dependency-free Python helper, bin/systemone-call, which posts standard System One request bodies to OpenCode Zen's jev-1.13-free model.
Jev Agents & MCP#github#codex#coding-agent
Communityaftely1337/typesafe-ai-opencode-free
A Claude Code plugin that sends every Agent spawn through JEV (TypeSafe), which picks the model (haiku, sonnet, opus), the effort and the tool profile of each subagent. It fails open to sonnet at medium effort, accepts a TypeSafe or OpenRouter key, and ships plan and run skills.
Jev Agents & MCP#community#github#claude-code
Community
hiyoku/jev-orcTypeScript
hiyoku/jev-orcTypeScriptAn experimental coding agent CLI built on opencode whose v2.0.0 architecture adds a JEV decision layer with five decision gates, shadow mode, and ECE calibration, plus a three-level RSI self-evolution loop and six Pi-framework extensions.
Jev Agents & MCP#community#github#typescript
Community
momozi1996/momo-codeTypeScript
momozi1996/momo-codeTypeScriptFrom 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.