jev-frontend-qa
Evidence-driven frontend QA where Jev picks browser actions while code checks DOM, HTTP and database contracts.
Overview
Jev Frontend QA is a command-line QA tool built on Jev Ultrafast and Browser Harness. Jev is the action selector only: it makes typed operation and target choices over observed controls, while exact fixture values and correctness checks stay in code. Contract runs check DOM state, correlated HTTP requests and responses, absent requests, and fresh database reads; exploratory runs never imply a contract pass. An explore mode lets an authenticated Claude or Codex CLI plan bounded subgoals while Jev still selects browser actions.
The README reports that live Jev acceptance passed on synthetic reference scenarios: healthy create, lifecycle and validation flows passed, and all three deliberate defects failed their contracts. Observations include accessibility names, form and dialog context, pagination, keyboard widgets and nested scrolling. Safety controls include a deny-by-default policy, redaction, bounded time and actions, and no automatic retry of ambiguous writes. A portable skill lets Claude Code or Codex author scenarios. It needs Python 3.12+, Chrome and SQLite for the demo.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which operation and observed control should the QA run use next? | — |
How it works
- Author a scenario (optionally via the coding-agent skill)
- Jev selects browser operations and targets over observed controls
- Code checks DOM, HTTP and database evidence against the contract
- Report pass or fail with screenshots, Jev usage and cost
Strengths
- Correctness judged by code, not the model
- Live acceptance with deliberate defects
- Detailed session accounting
Caveats
- No offline model fallback in the production CLI
- Reference scenarios are synthetic
At a glance
- Type
- App
- Status
- beta
- Language
- Python
- Runtime
- CLI with Browser Harness and Chrome
- Has code
- yes
- Has tests
- yes
Use cases and integrations
How curated lists describe it
Frontend QA that uses Jev to choose browser actions and checks contracts through DOM, HTTP and database evidence.
logicrw/awesome-jev-projectsEvidence-driven frontend QA built on Jev Ultrafast and Browser Harness, with a synthetic todo demo.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Which browser action to take next; it does not judge correctness.
How is correctness checked?
Code checks DOM, HTTP and database evidence against authored contracts.
Has it been tested live?
Yes, on synthetic reference scenarios including deliberate defects.
Related projects
Browser QA where Playwright drives and records while Jev judges assertions, matches controls to intent and picks goal-directed clicks.
Agent skillbetaTypeScriptPlaywright TypeScript project using Jev for semantic assertions, element picking, goal-driven steps and test failure classification.
ProjectprototypeTypeScriptPlain-English Playwright end-to-end tests where Jev resolves each step's target element and judges whether verify claims hold.
CLIprototypeTypeScriptLocal desktop pal that verifies what a coding agent really did, with optional Laya or Jev double-checks of unclear test results.
ProjectbetaTypeScriptDrop-in proxy for Playwright MCP that adds Jev page triage, prompt-injection masking, snapshot pruning and risky-action gating.
MCP serverbetaTypeScriptCodexQA browser skill that indexes in-page controls and has Jev pick the operation and target each step, with no screenshots needed.
Agent skillbeta