jev-browser-skill
Agent skill and local jb command that drive an isolated Playwright Chromium, with Jev choosing in-page actions for a narrow goal.
Overview
jev-browser-skill by hqman provides an agent skill named jb-browser and a local jb command that run an isolated Playwright Chromium. A coding agent or user sets a narrowly scoped browser goal, such as find the More information link and stop when that page is visible, and Jev chooses the in-page actions within a step limit. A daemon starts automatically on run, with commands to inspect state, stop and shut down, and artifacts stored under ~/.jb/data.
The default provider is the Vercel AI Gateway, with TypeSafe's System One API as an option, configured in ~/.jb/config.json or environment variables. It requires Node 22.18+ for TypeScript type stripping. It is distributed only as a GitHub clone, not an npm package (the README notes unrelated npm packages share similar names), and the skill folder must stay in the clone, linked into Cursor or other agent skill directories by symlink.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which in-page action should be taken next toward the narrow goal?(inferred) | — |
How it works
- Agent runs jb run with a URL, goal and max steps
- Daemon opens an isolated Chromium
- Jev chooses actions until the goal is visible
- Inspect state; stop or shut down
Strengths
- Isolated browser, separate from your own profile
- Works with Vercel AI Gateway or TypeSafe
- Simple CLI for agents
Caveats
- Not on npm; must be cloned
- Requires recent Node 22.18+
Getting started
`git clone git@github.com:hqman/jev-browser-skill.git && cd jev-browser-skill && npm install && npx playwright install chromium`, then `./bin/jb run --url ... --goal ...`.At a glance
- Type
- Agent skill
- Status
- prototype
- Language
- TypeScript
- Runtime
- agent skill, local daemon (Playwright)
- Has code
- yes
Use cases and integrations
How curated lists describe it
Agent skill that drives an isolated Playwright Chromium: the agent sets a narrow goal and Jev chooses the in-page actions. Vercel AI Gateway by default, TypeSafe optional.
MrJev/awesome-jev
Frequently asked questions
What does Jev decide?
The in-page actions needed to reach a narrow goal.
Which provider is used?
Vercel AI Gateway by default, or TypeSafe.
Is it on npm?
No, it is distributed as a GitHub clone only.
Related projects
MCP server and libraries that drive Chrome over CDP and hand a decision model like Jev a bounded menu of page actions.
ProjectproductionTypeScriptJev browser agent that pierces shadow DOM and iframes, adds canvas fallback and full actions, with one Jev request per cycle.
MCP serverbetaPythonMCP server that takes a whole browser task (URL, goal, checks) in one call and drives Chrome server-side with a decision model.
MCP serverbetaPython- Jev Browser★ 335
Headless browser automation via MCP server, CLI or library: Jev picks one action per step and scores goal-met and stuck probabilities.
MCP serverbetaTypeScript Claude Desktop MCP extension that drives Brave or Chrome with Jev choosing on-page actions and a guard freezing irreversible clicks.
MCP serverbetaTypeScriptChrome MCP server with 50 tools that drives your everyday browser without losing logins, where Jev or a local decider picks each step.
MCP serverbeta