Skip to content

jev-organize

Zero-dependency Node CLI that classifies a folder of company files with Jev by department, type, sensitivity, date, counterparty and PII.

Agent skillbeta

Overview

jev-organize takes a messy folder of company files (Word, Excel, PowerPoint, PDF, email, CSV, HTML, code and text) and classifies each one with TypeSafe's Jev on OpenRouter. For every file it decides department, document type and sensitivity with Choice questions; finds the document's own date and counterparty by having code extract candidates and Jev pick one; flags personal data, credentials and payment details with Nouls backed by regex detectors such as Luhn and IBAN checks; applies tags like draft or action-required; and detects duplicates and noise.

Sensitivity has a floor set by detectors that Jev cannot talk down. Output is a catalog, an INDEX.md map for people and AI agents, an offline HTML report and, on request, an organised copy with _Restricted, _Needs review, _Duplicates and _Noise folders. Originals are never moved or edited. It ships with a Claude skill and a Codex custom agent. The README cites about 0.4 s per file and 17 cents per 1,000 files.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich department does this file belong to?finance · legal · HR · sales · marketing · product · operations · IT · support · management
ChoiceHow sensitive is this file?public · internal · confidential · restricted
ChoiceWhich of the found dates is the document's own date?—
Yes / noDoes this file contain personal data, credentials or payment details?yes · no

How it works

  1. Read every file in the folder and extract text
  2. Code finds dates, company names and PII-shaped patterns
  3. One Jev call per file answers Choice and Noul questions
  4. Detectors enforce a sensitivity floor; low confidence goes to Needs review
  5. Write catalog, index, report and an optional organised copy

Strengths

  • Originals are never modified
  • Detector floors prevent under-classifying sensitive files
  • Index and AGENTS.md help AI agents work with the result
  • Low cost per file as reported

Caveats

  • File contents are sent to the provider via OpenRouter
  • Scans without text go to review
  • Requires Node 22.9+ and an OpenRouter key

Getting started

Requires Node 22.9+ and an OpenRouter key; see the Quick start section of the README.

At a glance

Type
Agent skill
Status
beta
Language
JavaScript
Runtime
Node CLI plus Claude skill and Codex agent
Has code
yes

Use cases and integrations

How curated lists describe it

  • Throw in a pile of company files and get them classified and organized by department, type, sensitivity, date, counterparty and PII, with an index for AI agents. Powered by TypeSafe's Jev on OpenRouter (17¢ per 1,000 files). Zero-dependency Node CLI + Claude skill + Codex agent.

    hellogumbo/awesome-jev

Frequently asked questions

What does Jev decide for each file?

Department, document type, sensitivity, which date and counterparty are the document's own, PII flags and tags.

Will it move my files?

No. Originals are never moved, renamed, edited or deleted; it writes an organised copy only on request.

How much does it cost?

The README cites about 17 cents per 1,000 files.

Related projects

  • Skill and plugin for Codex, Claude Code and other agents that helps them use Jev for bounded judgments and measure the effect.

    Agent skillbetaJavaScript
  • MIT plugin and local MCP server that hands an agent's closed questions to Jev or local Laya, with policy gates and receipts.

    Agent skillproductionPython
  • Claude Code and Codex skill that routes a task to one of 27 specialist roles, with an optional Jev engine for catalog selection.

    ProjectbetaPython
  • Orchestrator that probes which coding agents are installed and logged in, then has Jev pick the best one for a task.

    Agent skillbetaPython
  • Hooks for Claude Code, Codex and pi that use Jev to suggest a skill each turn, plus a rule gate for subagent model routing.

    HookbetaJavaScript
  • MCP server and CLI where Jev decides if a coding task needs a Conductor cloud workspace, how hard it is, and which harness and model run it.

    ProjectbetaTypeScript