Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

Agent Tool

IDE and browser extensions to find and install agent Skills and Subagents, with an opt-in Jev check when local detection can't decide.

Browser extensionproductionMIT
View on GitHub ↗

Overview

Agent Tool has two parts sharing detection and install logic: an IDE extension for Cursor and VS Code that shows Skills, Subagents, Rules, MCP servers and Plugins across coding agents in one dashboard, and a Chrome, Edge and Brave extension that spots Skills and Subagents while you browse and installs them into folders you pick.

Jev plays a small, optional role. On unsupported sites, detection runs locally in the page; as a beta feature, when local evidence cannot decide, you can add your own Jev API key so a single question goes to TypeSafe. That is off by default, supported sites stay deterministic and never call Jev, and the README warns the experiment may change or be withdrawn.

What the decision model decides

TypeQuestionAnswers
Yes / noDoes this page contain a GitHub-backed Skill or Subagent?(inferred)yes · no

How it works

  1. Browser extension scans the current page locally
  2. On supported sites, detection is deterministic
  3. On unsupported sites with no clear local evidence, optionally ask Jev one question
  4. Offer to install into a chosen folder

Strengths

  • Jev is strictly opt-in with your own key
  • Works across macOS, Linux and Windows

Caveats

  • Jev feature is beta and may be withdrawn
  • Decision model plays a minor role

Getting started

Install from Open VSX (yuuki-sakai/agent-tool) for Cursor/VS Code, or the browser extension for Chrome/Edge/Brave.

At a glance

Type
Browser extension
Status
production
Language
TypeScript
Runtime
VS Code/Cursor extension, browser extension
Has code
yes
Has tests
yes
License
MIT

Use cases and integrations

How curated lists describe it

  • Cursor/VS Code extension and a Chrome/Edge/Brave extension that catch Skills and Subagents while you browse and install them into your coding agents.

    hellogumbo/awesome-jev

Frequently asked questions

When is Jev used?

Only on unsupported sites, when local detection cannot decide and you have enabled it with your own key.

What does it install?

Skills and Subagents, plus Claude Code and Codex Plugins from the IDE extension.

Is Jev on by default?

No, it is off until you enable it.

Related projects

  • bf-jev-deep-research

    Portuguese agent skill teaching Claude Code, Cursor and Codex to use Jev correctly, bundled with the full raw research behind each claim.

    Agent skillproduction
  • agent-fastpath

    MCP server giving coding agents fast typed decisions (ship gates, risk checks, file triage, page checks) as accept, review or escalate.

    MCP serverbetaTypeScript
  • jev-awesome-skills

    Open-source Jev skills for Claude Code, Codex, Cursor and Grok: one client to gate, route, guard, verify, triage, review and act.

    ProjectbetaTypeScript
  • jev-classifier

    Local gateway and MCP tools that let Jev predict which tool a coding agent should call next, from suggestions to enforced choices.

    MCP serverbetaTypeScript
  • jev-max

    Jev browser agent that pierces shadow DOM and iframes, adds canvas fallback and full actions, with one Jev request per cycle.

    MCP serverbetaPython
  • jev-ultrafast-mcp

    MCP server that takes a whole browser task (URL, goal, checks) in one call and drives Chrome server-side with a decision model.

    MCP serverbetaPython