User repository on GitHub
Jev picks which of your rules apply to each prompt, so Claude only sees the ones that matter.
- Forks
- 11
- Last pushed
- Updated Sep 2026
Repo stats from the GitHub API, cached Oct 2026.
CommunityJev Agents & MCP
Jev Rules
Claude Code plugin that asks Jev one yes/no question per rule and passes Claude only the matches, so large rule folders stay out of context until relevant.
- Stars
- 66
- Language
- JavaScript
- License
- MIT
- Published by
- Community
- Author
- EliaAlberti
- Last verified
- Oct 2026
About this project
A real Claude Code session with the rules pane, edited to 30 seconds. The full-quality video is attached to the v0.5.0 release . The earlier 40-second demo is still online.
Quoted from the project's README. Read the full README.
Works forJev for Classification
Used inJev for Software & DevTools
- Languages
- JavaScript
Highlights
- Rules live as Markdown in .claude/jev-rules/ with a one-sentence description Jev judges, plus optional always, applies, and does_not_apply fields.
- Matching rules and filtered codebase-map documents from .claude/jev-map/ are delivered once per session, with file-following rules and subfolders.
- Installs with plugin marketplace commands and a Jev key file in the home directory, with no dependencies beyond Claude Code and Node.js.
- Falls back to showing every rule when the decision call fails, and retries when the API is busy, so no rule is silently lost.
Watch out
Needs a TypeSafe Jev key and Claude Code with plugin support; rule quality depends on concrete one-line descriptions. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
Keyword matching misses paraphrases and full loading wastes context; Jev judges each description per prompt and only matching rules reach Claude, with once-per-session delivery capping long-session cost.
A miss means Claude never sees that rule for the prompt, so descriptions must name files, actions, and areas concretely; on errors it falls back to showing every rule and retries busy API responses.
Use one plain sentence answering what the request is about, such as payments or deploys, and add applies or does_not_apply only when a description keeps catching the wrong prompts.
Claude Code loads its own rules folder at session start, which is what this avoids; the separate jev-rules folder plus the jev-map folder lets Jev filter rules and map documents per prompt.
Inside Claude Code run /plugin marketplace add EliaAlberti/jev-rules, then /plugin install jev-rules@jev-rules, and /reload-plugins or restart if you installed mid-session. Put JEV_API_KEY=your-key in ~/.jev-rules.env; a Vercel AI Gateway key works instead as AI_GATEWAY_API_KEY. You need Claude Code with plugin support and Node.js 20.12 or newer on PATH.
Rules tell Claude how to work; map documents in .claude/jev-map/ tell it how the project fits together, such as where the checkout code lives or how a release goes out. Each document carries the same header and is filtered by the same decision, so a prompt gets the one document that answers it instead of the whole map. always is ignored there, and a document with no description falls back to its title and first paragraph, up to 300 characters.
As of 0.5.0, one line under each prompt names the rules Claude was given with Jev's score for each, and the rules pane shows every rule green once Claude has it; Claude asks once whether to switch the pane on the first time Jev picks a rule. Set JEV_DEBUG=1 for a ~/.jev-rules.log. To try a working copy without installing, run claude --plugin-dir pointing at plugins/jev-rules.
More Jev resources like this
47GitHub stars
MCP server that screens files, public webpages, or inline text against a query and returns compact relevance probabilities, so the main agent opens only promising items.
Jev Agents & MCP#community#jevbest#catalog-import
Community
kbhuw/jev-siftJavaScript
kbhuw/jev-siftJavaScript16GitHub stars
Standalone terminal agent that puts Jev inside its planning loop, proposing text-model steps and scoring candidates against independent rubrics before a deterministic policy executes one.
Jev Agents & MCP#community#jevbest#catalog-import
19GitHub stars
Capability resolver that uses Jev to search a nested command catalog and return selected instruction context to an agent through one MCP tool, keeping the search outside agent context.
Jev Agents & MCP#community#jevbest#catalog-import
Community
NiazMorshed2007/jcrJavaScript
NiazMorshed2007/jcrJavaScript7GitHub stars
MCP server exposing one evaluate tool for Jev Choice, Noul, and Score questions, plus a separate eval harness that compares Jev against LLMs on shared labeled sets.
Jev Agents & MCP#community#jevbest#catalog-import
Community
arunav25/jev-mcpJavaScript
arunav25/jev-mcpJavaScript9GitHub stars
MCP tool router for Codex that keeps large external tool inventories behind a small find_tool surface and asks Jev to select from names and descriptions at a 0.90 threshold, falling back to bounded search.
Jev Agents & MCP#community#jevbest#catalog-import
Community
jackbarunz/jev-tool-routerJavaScript
jackbarunz/jev-tool-routerJavaScript3GitHub stars
Proxy that sits between any MCP client and any stdio MCP server and screens tool calls, tool results and tool descriptions with Jev probabilities, layering deterministic allowlists, semantic checks and session taint tracking over a JSONL audit trail.
Jev Agents & MCP#community#jevbest#catalog-import
Community
caiovicentino/jev-shieldJavaScript
caiovicentino/jev-shieldJavaScriptFrom the Jev community
Posts from builders shipping with Jev right now.