Do I need an AI model just to check robot health?

No. The REPL, deterministic tests and SQLite history work without any model. AI only adds natural-language diagnosis over the same read-only tools when an OpenAI-compatible endpoint or Jev key is set.

Why does battery show N/A instead of FAIL on Mini Pupper 2?

RDCD declares what a robot should expose. Mini Pupper 2 omits battery, so battery shows as N/A meaning not supported rather than a fault, and it does not lower system health.

Can the AI drive the robot or bypass safety checks?

No. Jev only decides which read-only evidence tool to call next and the LLM only explains collected evidence. Motion permission stays with the deterministic Safety Gate code.

What robot is actually supported right now?

Mini Pupper 2 is the first reference with diagnostics, joint states, IMU, odometry and velocity command. Others run in generic discovery mode or with a custom YAML RDCD file.

What two jobs does Jev do in the diagnostic routing?

It classifies each request as QUERY, DIAGNOSIS, ACTION, or CHAT, and for a diagnosis it chooses which read-only evidence tool to call next. The LLM writes the final explanation, and Jev does not execute motion or write operations.

What does the /stop command actually do?

It sends a software stop through the configured velocity topic and can optionally call a std_srvs/Trigger service. The README notes a software stop is not a replacement for a physical emergency stop.

What happens when an expected capability is missing at runtime?

It stays a real diagnostic finding and the Safety Gate fails closed. That differs from a capability the RDCD never declared, which is reported as N/A and does not reduce the overall system health result.

CommunityJev Demos & Experiments

Robodiag Harness

Terminal harness that checks ROS 2 robots with deterministic health tests, capability descriptions, SQLite history, and optional Jev evidence routing.

See highlights

Open github.com

Stars: 22
Language: Python
License: APACHE-2.0
Published by: Community
Author: YueBit
Last verified: Oct 2026

YueBit/robodiag-harness

User repository on GitHub

View on GitHub

AI diagnostic agent for ROS 2 robots — evidence-based, tool-calling, safety-gated.

Forks: 0
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

About this project

A diagnostic harness for ROS 2 robots. Your robot is connected. Is it actually healthy? Is telemetry arriving at the expected rate? Are joints reporting valid values? Are controllers active? What evidence should you collect when something goes wrong? RoboDiag brings robot telemetry, deterministic health checks, diagnostic history and AI-assisted investigation into one terminal interface.

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

Tagscommunity jevbest catalog-import

Works forJev for Classification Jev for Moderation Jev for Tool calling

Used inJev for Software & DevTools

Languages: Python

Highlights

Robot to evidence to tests to diagnosis flow with the live ROS graph, diagnostics and topic sampling.
RDCD capability files declare expected interfaces so missing data is a finding and undeclared is N/A.
Deterministic tests for graph, diagnostics, battery, joints and controllers with PASS, WARN, FAIL and SKIP.
Safety Gate and software stop live in deterministic code, separate from diagnostic reasoning.

Watch out

Needs ROS 2 with matching domain and sourced workspaces, and software stop is not a physical emergency stop. 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-style

Local Qwen3.5 decision models in 0.8B and 2B sizes with fitted temperatures, systemone-shaped server, playground, guard, MCP tools, and eval harness.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

lawrence3699/jev-stylePython

Jev_Star

StarCraft II research repo combining Jev action selection with optional Astra plans across full-game macro control and SMAC-Hard micromanagement, with videos, replays, and a paper.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

sc2musa/Jev_StarPython

jev-kit

Claude Code kit that guards PreToolUse calls with a free pre-filter plus one Jev question for ambiguous cases, and adds sub-agent sizing, rechecks, search, daemon speedups, and monitoring.

Jev Demos & Experiments#community#awesomejev#catalog-import

Community

jonathanavis96/jev-kitPython

jev-ultrafast-mcp

MCP server where one browser_goal call hands a URL, a goal, and code-checked assertions to a server-side loop driven by a decision model over live page elements.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

jiawei686/jev-ultrafast-mcpPython

open-spark-jev

Independent open-source local decision model inspired by Jev, with a spark-s1 family plus smaller control models for agent gates, served locally with probabilities and policy actions.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

abhishek085/open-spark-jevPython

fastbrowse

114

Browser agent where Jev picks each action from indexed page candidates while an LLM plans and reads, returning cited answers or schema-checked data with credential controls.

Jev Demos & Experiments#community#jevbest#catalog-import

Community

agent-labs-dev/fastbrowsePython

From the Jev guides

Original write-ups on what this entry is for.

Per-Tick Decisions in Games

Sep 28, 2026

A game tick is a state plus a legal action list. How Clash Royale, Minecraft, and the earlier Mario, Pokémon, StarCraft, and drone loops ask Jev one step at a time.

Jev Demos & Experiments#gaming#agents#control

Read article

Back to all resources

From the Jev community

Posts from builders shipping with Jev right now.

Follow @typesafeai