jevify
Agent skill that teaches coding agents to hand bulk semantic judgments to Jev and to find where a product could use Jev decisions.
Overview
Jevify is an agent skill that gives an agent a System One capability. An agent can have access to Jev and still read every document, load every tool result and repeat every judgment itself; Jevify teaches it to recognise work a fast, inexpensive decision model can handle and gives it the tools to act, such as selecting useful passages, ranking candidates, checking many items or choosing a handler, then bringing selected evidence back into context to finish the task.
There are two entry points. During ordinary tasks, a large document, noisy tool output or repeated semantic checks can trigger Jevify even if you never mention Jev, within the available runtime and external-inference permissions; the skill makes Jev the default for bulk semantic judgments rather than first building a cost comparison. Asked to jevify a product or workflow, it inspects it and identifies concrete improvements, including capabilities that frequent cheap judgments could enable, and designs comparison experiments to measure whether Jev helped. The skill covers Choice, Noul and Score, shared-state batching and question design, and learns from recent community examples. MIT licensed.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Score | Which passages or candidates are most relevant to the task?(inferred) | — |
| Yes / no | Does each item pass this semantic check?(inferred) | — |
| Choice | Which handler should process this item?(inferred) | — |
How it works
- Agent detects bulk semantic work or is asked to jevify
- Delegate the first pass to Jev with typed batched questions
- Read selected evidence and continue reasoning
- For products, propose decision points and comparison experiments
Strengths
- Reduces agent reading and reasoning tokens on bulk judgments
- Finds product opportunities for Jev
- Draws on recent community examples
Caveats
- Assumes Jev is cheaper without measuring each time
- Requires permission for external inference
At a glance
- Type
- Agent skill
- Status
- beta
- Runtime
- Agent skill
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Agent skill: finds where a codebase could hand a decision to Jev, designs the typed questions for it, and learns from recent community usage.
yibie/awesome-jevAn Agent Skill for finding suitable Jev decision points and designing questions and comparison experiments.
logicrw/awesome-jev-projectsAn agent skill to discover TypeSafe Jev opportunities, design typed questions, and learn from recent community experiments.
hellogumbo/awesome-jevAgent skill for finding where Jev fits in an existing codebase, designing the typed questions, and measuring whether it helped.
MrJev/awesome-jev
Frequently asked questions
When does the skill activate?
On large documents, noisy tool output or repeated semantic checks, or when asked to jevify something.
Does it replace the agent's reasoning?
No, Jev handles the first-pass judgments and the agent finishes the task.
Can it suggest product improvements?
Yes, it identifies concrete decision points and designs experiments.
Related projects
- Building with Jev★ 155
Agent skill for writing and improving programs that call Jev: question design, state structure, thresholds and diagnosing wrong answers.
Agent skillproduction pip-installable Jev kit: a CLI for verify, screen, classify, extract, route and rerank, plus an MCP server for the primitives.
MCP serverbetaPythonPortable Agent Skill teaching coding agents to use Jev as a bounded judgment layer, plus an optional MCP bridge.
MCP serverbetaTypeScriptIndependent agent skill that makes a coding agent design, build, test and tune a bounded Jev decision layer with a contract first.
Agent skillbetaAgent skill that guides you in finding, designing and testing Jev decision loops, including when to keep logic in code.
Agent skillbetaMCP server giving coding agents typed judgment tools (verify, screen, rerank, gate, score) backed by Jev, with auto/review/escalate policy.
MCP serverbetaPython