What problem does Jevflow solve?

AI coding agents declare victory early, drift from the goal, trip over each other in one repo, or stop mid-run after a crash or rate limit. Jevflow keeps Claude accountable by running checks and asking Jev where the work really stands.

How does a flow work?

For any real task, Claude lays the work out as a small plan of phases, each with a definition of done and a check. Every time Claude tries to stop, Jevflow runs the checks and asks Jev whether a phase is finished, whether Claude is stuck, off the goal, or claiming done too early.

Who decides what happens next?

A small, fully tested policy, not the model: it returns keep going, move to the next phase, or you are really done. The README's badge reports 375 passing tests for the plugin.

Can several agents work together?

Yes. Several agents can share one plan, every flow in a project is remembered, and a live viewer shows it all; the demo shows three agents working three parallel phases of one flow before the test loop, release tag and goal complete.

You need Claude Code, Python 3.10+ and a Jev API key from TypeSafe. Add the marketplace with claude plugin marketplace add Parth1811/JevFlow, then claude plugin install jevflow@jevflow; the same commands work as /plugin commands inside Claude Code.

What if Python or the API key is missing?

Without a Jev API key Jevflow stays off and says so once per session. macOS ships Python 3.9, so install a newer one first; if Claude still cannot find it, set JEVFLOW_PYTHON to its full path.

What is the project status?

MIT licensed and labelled MVP in the README badges. The plugin has no runtime dependencies beyond the Python standard library, and it is distributed as a Claude Code plugin through the Parth1811/JevFlow marketplace.

Jev directory

CommunityJev Agents & MCP

Jevflow — parth1811

A Claude Code plugin that keeps agents honest: Claude lays work out as phases with a definition of done and a check, and whenever it tries to stop, Jevflow runs the checks and asks Jev whether the work is really done. Several agents can share one plan.

See highlights Open github.com

Stars: 10
Language: Python
License: MIT
Reading time: 2 min
Published by: Community
Author: parth1811
Last verified: Oct 2026

parth1811/jevflow

User repository on GitHub

View on GitHub

A Claude Code plugin that keeps your AI agents honest. Claude plans tasks as phases with checks, and when it tries to stop, Jevflow runs them and asks Jev if the work is really done. Many agents, one plan, with a live viewer to watch it all.

Forks: 1
Last pushed: Updated Oct 2026

Repo stats from the GitHub API, cached Oct 2026.

This is one of several Jev listings for the same project. The canonical entry is Jevflow — Mawfyy.

Tagscommunity catalog-import logicrw claude-code plugin multi-agent

Works forJev for Verification Jev for Evaluation Jev for Decision gating

Used inJev for Software & DevTools

Languages: Python
Platforms: CLI
Capabilities: Open sourceRequires API key

Highlights

On every stop attempt, Jevflow runs the phase checks and asks Jev if the work is finished, stuck or off the goal.
A small, fully tested policy decides keep going, move to the next phase, or you are really done.
Ships as a Claude Code plugin from the Parth1811/JevFlow marketplace, with a live viewer via jevflow ui.
Multiple agents and subagents can share one plan; every flow in a project is remembered.
Python 3.10+, stdlib only, 375 passing tests, MIT, status MVP; a Jev API key is required.

Quickstart

bash

claude plugin marketplace add Parth1811/JevFlow
claude plugin install jevflow@jevflow

Watch out

Grounded on a truncated README and GitHub metadata captured 2026-10-15; the digest cuts off during the API-key setup step, so key configuration beyond 'Claude Code asks for the key' is taken from the plugin description. macOS ships Python 3.9, so a newer Python must be installed first.

Questions about this entry

Short answers drawn from the linked project or page.

Reactions & coverage

Posts, threads, and videos about this entry from around the web.

YouTube: Ray Amjad puts Jev inside an agentic coding loop and looks at what it costs to run.

YouTube: RoboNuggets puts Jev to work inside Claude Code.

Reddit: A Claude Code mod replaces the auto-mode classifier with Jev, 2x faster.

From the Jev guides

Original write-ups on what this entry is for.

MCP tool routing: Jev in front of the tool call

Oct 16, 2026

Tool menus can outgrow the context that reads them. The catalog's Jev-based routers and gates narrow the list: benchmarked tool search, per-tool scoring, fail-closed approvals, and WebMCP in the browser.

Jev Agents & MCP#mcp#tool-routing#agents

Read article

Approval policies for coding agents: hooks, guards, and gateways

Oct 16, 2026

How the Jev guard family turns coding-agent permissions into auto-approve, deny, or ask: hooks that enforce, guards that fail safe, and gateways that gate generation.

Jev Agents & MCP#coding-agents#guardrails#hooks