Can I run the refund or host-isolation paths unattended?

No. Thresholds, procedures, and labels are not claimed production-ready, and Jev never sends mail, moves money, or runs shell commands. A flat probability distribution means the option descriptions do not separate the options.

Why did my extra context fields make accuracy worse?

Accuracy falls when state carries fields the question does not need, and contradictory criteria score worse. User-controlled text can change the answer, so keep state narrow and let code compute sums, dates, and duplicate flags.

Are the cost and latency figures independently verified?

Those figures are TypeSafe's launch numbers, not an independent benchmark: input-only billing with parallel evaluation in one request and schema match guaranteed. Jev can still return the wrong valid key, so threshold on confidence in code.

What does the SOC pipeline actually execute?

Triage, investigation, mitigation, and escalation run in order and only name actions like collect and page candidates. Mitigation approves or holds isolate, revoke, reset, and block actions without performing them.

How do I run the use-case modules against the live API?

Each module builds typed state and Jev questions, calls the live TypeSafe API, applies decision logic in code and returns a UseCaseResult. On 18 September 2026 the 27 Jev-only runners each returned a decision from jev-1.13.0 on the committed fixtures.

What fields does a UseCaseResult carry?

A decision, an action band of auto, confirm, human or block, and the actions the caller performs. Jev does not send email, move money or run a shell command.

Which decision types does the repository implement?

Classification, detection, scoring, routing, search and ranking, verification, feature extraction, and bounded extraction, spread across routing, guardrails, records and risk, and people and research modules.

CommunityJev Agents & MCP

Jev Usecases

Callable Python collection implementing most documented Jev decisions, where each module builds state, asks typed questions, and returns a UseCaseResult for code to execute.

See highlights

Open github.com

Stars: 22
Language: Python
License: MIT
Published by: Community
Author: kenhuangus
Last verified: Oct 2026

kenhuangus/jev-usecases

User repository on GitHub

View on GitHub

Production TypeSafe Jev (System One) use-case harnesses with confidence-gated decision logic

Forks: 5
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Jev returns Choice, Score, and Noul probabilities in 70 to 500 ms. This repository is the initial implementation of most of the decisions TypeSafe lists for that model.

Quoted from the project's README. Read the full README.

Tagscommunity awesomejev catalog-import

Works forJev for Classification

Used inJev for Software & DevTools

Languages: Python

Highlights

Covers routing, verification, records-and-risk, and people-catalog-research modules plus SOC runners.
Returns decision, action band, and actions with auto, confirm, human, or block bands set by the caller.
Calls one endpoint with state plus Choice, Score, and Noul questions evaluated in parallel.
Documents jaggedness: literal reading, weak counting and date comparison, and harm from unneeded fields.

Watch out

Initial implementation whose fixtures only check that each runner returns a typed answer, not correctness; set confidence cutoffs on labeled traffic before automatic action. 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.

More Jev resources like this

jev-recruiter

Local LinkedIn recruiting workspace on Browser Use's Jev Ultrafast: Jev screens titles, drives navigation, and assesses profile excerpts against your brief while a live browser stays visible; matches and evidence export as JSON.

Jev Agents & MCP#community#jevbest#catalog-import

Community

skeptrunedev/jev-recruiterPython

jev-skill-suggester

Codex Skill plus stdlib-only Python CLI that recommends which installed skill to read next for a task, allowing no-skill or uncertain answers and never executing candidates.

Jev Agents & MCP#community#jevbest#catalog-import

Community

win4r/jev-skill-suggesterPython

jev-judge-mcp

MCP server over stdio that gives coding agents fourteen Jev judgment tools, turning evidence plus an enumerable question into auto, review, or escalate actions from calibrated probabilities.

Jev Agents & MCP#community#jevbest#catalog-import

Community

PyModel/jev-judge-mcpPython

jev-agent-design-with-topk-logits-choices

Research runtime that pages a large logical tool and context space through a small bounded Jev decision surface with explicit approvals.

Jev Agents & MCP#community#jevbest#catalog-import

Community

6Mikao9/jev-agent-design-with-topk-logits-choicesPython

muse-jev-playbook

Agent-oriented playbook for using Jev as a cheap, fast decision layer that triages, classifies, scores, and gates work before expensive browser, research, retry, or subagent steps.

Jev Agents & MCP#community#awesomejev#catalog-import

Community

Bodila51/muse-jev-playbookPython

claude-jev

Coding-agent plugin that routes prompt classification, subagent tiering, edit rule checks, and fast Jev-backed compaction through small typed judgments.

Jev Agents & MCP#community#jevbest#catalog-import

Community

0x7067/claude-jevTypeScript

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai