jevelry
Tool that lets code and agents ask Jev typed questions via reusable jevels, mapping answers to act, mark or fall_back.
Overview
jevelry lets your code and your coding agent ask Jev a small typed question with one command and get back a decision they can act on, while keeping every decision so you can later see how often Jev was right. Questions live in small files called jevels, and seventeen ready-made jevels ship with the package covering tickets, messages, issues, pull requests, logs, alerts and failing tests: is this ticket urgent, which team should get it, is this bug a duplicate, did this test fail because of the code or the machine.
Each jevel turns Jev's confidence into one of three decisions: act when Jev is sure, mark when it is fairly sure (use the answer but flag it for a person), and fall_back when it is unsure (do whatever the code did before). An interactive full-screen terminal UI built with Ink supports browsing and tracking, and keyboard shortcuts and configuration are documented. The README warns that it is in early development and that commands, the jevel format and the answer document may change before 1.0.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Is this ticket urgent? | — |
| Choice | Which team should get this ticket? | — |
| Yes / no | Is this bug a duplicate? | — |
| Choice | Did this test fail because of the code or the machine it ran on? | code · machine |
How it works
- Pick or write a jevel
- Ask Jev from code, an agent or the CLI
- Map confidence to act, mark or fall_back
- Record decisions and review accuracy over time
Strengths
- Seventeen ready-made jevels
- Three-way outcome with explicit fallback
- Decision history to measure accuracy
Caveats
- Early development; formats may change before 1.0
- Requires a TypeSafe key
At a glance
- Type
- Project
- Status
- prototype
- Language
- TypeScript
- Runtime
- Node CLI / TUI
- Has code
- yes
Use cases and integrations
How curated lists describe it
Use Jev everywhere to make & track decisions.
hellogumbo/awesome-jevPuts Jev behind ordinary decisions across a codebase and keeps a record of each one it makes.
MrJev/awesome-jev
Frequently asked questions
What is a jevel?
A small file holding one typed question and how to map Jev's confidence to a decision.
What are the three decisions?
act (Jev is sure), mark (fairly sure, flag for a person) and fall_back (unsure, keep previous behaviour).
Is it stable?
No, it is in early development and formats may change before 1.0.
Related projects
TypeScript library adding policy, confidence gates, shadow mode, recipes and an eval CLI around Jev answers.
CLIbetaTypeScriptClaude Code skill that hands bulk 'read a lot to decide a little' judgments over files, logs and tickets to Jev and returns a shortlist.
Agent skillbetaJavaScriptRust crate and jud CLI where a decision is a .jud rubric file answered by a System One model, type-checked and testable without a key.
PluginbetaRustRunnable Jev cookbooks, a plain-English guide and sixteen interactive simulations of how a typed decision model changes architecture.
AppbetaPythonGitHub Action that triages issues with Jev, applying label, spam, needs-info and duplicate tags only above set thresholds.
ProjectbetaTypeScriptOne YAML task to classify, route, score or verify text with rules, Jev, NanoJev or an OpenAI-compatible model, with cost records.
CLIbetaPython