JevScout
Coding-agent skill that drives Chrome over CDP to hunt for jobs on company sites, with Jev scoring every link and job posting.
Overview
JevScout (Jev AI Job Hunter) is a demo MVP skill for coding agents that searches real company websites for AI and software jobs. Its rule is 'Chrome sees and acts, Jev decides': a CLI called jjh drives a visible Chrome over the DevTools protocol, and Jev scores every link to decide where to click and every job page to decide whether it matches the query. The host agent's LLM never chooses what to click.
A single command starts from any company homepage and a query, for example an applied AI engineer role, and logs the results. A mock mode works without an API key, and a slow-motion mode steps through recorded page fixtures for narrated demos. Tests run with pytest. It illustrates how to keep a browsing loop's navigation decisions in a cheap typed model while the agent only invokes the skill.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Score | How likely is this link to lead to relevant job listings?(inferred) | — |
| Yes / no | Does this job posting match the query?(inferred) | — |
How it works
- Agent runs jjh run with a company URL and query
- Chrome loads the page; jjh extracts links
- Jev scores links and the best one is followed
- Job pages are scored against the query and matches saved
- jjh log shows results
Strengths
- Host LLM does not spend tokens on navigation
- Mock mode and fixtures for testing without a key
- Works from any company homepage
Caveats
- Demo MVP
- Requires visible Chrome with CDP
- Site structures vary widely
Getting started
cp .env.example .env (set TYPESAFE_API_KEY, optional with --mock), uv sync, then uv run jjh run --url https://openai.com --query "Applied AI Engineer, Codex".At a glance
- Type
- Agent skill
- Status
- prototype
- Language
- Python
- Runtime
- CLI and agent skill driving Chrome via CDP
- Has code
- yes
- Has tests
- yes
Use cases and integrations
How curated lists describe it
A demo job-search Skill for coding Agents that browses company careers pages in Chrome, uses Jev to screen AI and software-engineering roles, and saves the results.
logicrw/awesome-jev-projectsCoding-agent skill that drives Chrome over CDP to look for jobs on company sites, with Jev scoring pages and links.
MrJev/awesome-jev
Frequently asked questions
What does Jev decide?
Which links to follow and whether each job page matches the search query.
Does the coding agent choose clicks?
No, the README says the host LLM never picks what to click.
Can I try it without a key?
Yes, with --mock and the included page fixtures.
Related projects
MCP server and libraries that drive Chrome over CDP and hand a decision model like Jev a bounded menu of page actions.
ProjectproductionTypeScript- 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 serverbetaStealth browser agent where Jev decides each step in about 0.3 s and CloakBrowser carries it out; MCP server, CLI and Python API.
MCP serverbetaPythonCodexQA browser skill that indexes in-page controls and has Jev pick the operation and target each step, with no screenshots needed.
Agent skillbeta