Does Claude massage the probabilities?

No. The plugin builds one deterministic state from the live table, sends fixed Noul, Choice, and Score questions to Jev, and formats the raw answers into tables. Claude never estimates or corrects the numbers.

Where does the table come from?

ValorFinal open data under a CC BY license, computed from football-data.org results. Only objective per-club fields like position, points, games, wins, draws, losses, and goal difference enter the state.

Why do Noul and Score disagree on one club?

That was observed in real runs and is preserved in the report without correction as model behavior to document, not a plugin bug. Each round persists the exact state, questions, raw replies, and timestamps for comparison.

What breaks on Claude Desktop?

The native Linux host bridge does not expand user-config secrets in local MCP servers and silently drops the jev server. A workaround script configures the desktop file for you, but it stores the key in plaintext locally, so use it for testing only.

How do I install the plugin and run a round?

Clone the repo, change into the folder holding .claude-plugin/plugin.json, and run ./scripts/package.sh, which installs dependencies, rebuilds the server, and produces an upload-ready zip. Add it through Customize, Plugins, Add, Upload plugin, enable it, then run /brasileirao. Zipping the folder by hand can pull in server/node_modules and break the upload.

What does one round of /brasileirao actually ask?

Forty-one questions over one deterministic state: twenty per-club Noul asking whether the club will be relegated, one Choice with the twenty clubs as options for the champion, and twenty per-club Score questions on a zero-to-five band from relegation to champion.

What stops a bad run before it reaches the model?

Step two of the six-step command validates the table first: twenty clubs, no duplicates, wins plus draws plus losses consistent with games played, and an identified round. Any failed step halts the run and reports the error rather than asking Jev about a broken state.

CommunityJev Integrations & Adapters

Brasileirao Jev

Claude Cowork plugin running reproducible Brasileirao Serie A probability experiments through Jev over Vercel AI Gateway, with Claude only fetching the table and presenting raw numbers.

See highlights

Open github.com

Stars: 1
Language: JavaScript
Published by: Community
Author: andretestanalyst
Last verified: Oct 2026

andretestanalyst/brasileirao-jev

User repository on GitHub

View on GitHub

Plugin para Claude Cowork que usa o modelo Jev (TypeSafe AI) via Vercel AI Gateway para análises probabilísticas do Brasileirão Série A

Forks: 0
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

Plugin para Claude Cowork que usa o modelo Jev (TypeSafe AI), via Vercel AI Gateway, para rodar experimentos probabilísticos (Noul, Choice, Score) sobre o estado atual do Campeonato Brasileiro Série A. A especificação original do projeto, com as decisões e limitações encontradas durante a implementação, está em docs/brasileiraojevcoworkplugin.md.

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

Tagscommunity github

Works forJev for Classification

Used inJev for Software & DevTools

Languages: JavaScript

Highlights

Deterministic state built from live ValorFinal standings, discarding that source's own projections.
Each run asks 20 per-club Noul relegation questions, one 20-option champion Choice, and 20 band Scores.
Six-step command validates the table, runs Jev, and persists raw JSON plus a report per round.
Gateway key stays in the platform vault as an env var; rejected keys fail without fabrication.

Watch out

Not a statistical simulator and makes no accuracy claim; Noul and Score for the same club can disagree and are kept as-is, and bands ignore cup-redistributed slots. 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-crawlers

Unix-style bug-discovery pipeline: small tools pass JSON lines through seed, expand, judge, verify, and report stages, spending one cheap Jev judgment per code node to chase leads across callers and configs.

Jev Integrations & Adapters#community#github

Community

Thingscorp/jev-crawlersJavaScript

dsh-opencli

DeepSeek Harness plugin that gives agents hands in the user's own logged-in browser: 176 site adapters with 200+ one-shot commands plus generic page primitives, with write actions behind an approval gate.

Jev Integrations & Adapters#community#github

Community

IKEASven69/dsh-opencliJavaScript

dsh-jev

DeepSeek Harness community plugin that classifies each turn's newest user message through the systemone endpoint and injects the verdict as tagged runtime context for tone and priority.

Jev Integrations & Adapters#community#github

Community

Excalibur9527/dsh-jevJavaScript

jev-gateway

Local HTTP service exposing Jev through the Vercel AI Gateway: projects POST state plus typed questions and get Jev answers with provider metadata, without holding a TypeSafe API key.

Jev Integrations & Adapters#community#github

Community

nitinnat/jev-gatewayJavaScript

ai-backends

146

A self-hostable API server that fronts text, vision and data endpoints across Ollama, LM Studio, LlamaCpp and a long list of hosted providers, and also runs tool-using agents with sessions. Jev appears separately as a decision provider for typed choice, score and noul questions.

Jev Integrations & Adapters#community#jevbest

Community

donvito/ai-backendsHTML

ST-jeved

SillyTavern extension that sends chat text to Jev sensors and fires one rule action per turn such as nudging, rerolling, editing lists, or running scripts.

Jev Integrations & Adapters#community#jevbest#catalog-import

Community

mossyfield/ST-jevedJavaScript

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai