
from agentic-test-runner8
Run natural-language browser behavior tests and AI-driven end-to-end checks with ATR; captures failures and provides debug info.
ATR Behavior Testing lets agents run browser-based behavior tests written in plain English (.test.txt). The skill interprets test files, executes actions via browser automation, and collects screenshots, console logs, network traces and DOM snapshots for failure analysis. It provides command-line flows to run single tests or directories, debug with a live browser, and adjust viewport or timeouts for reliable automation.
Use this skill when you want to validate user journeys (login, checkout, form submissions) without writing low-level selectors. It's ideal for QA automation, regression checks after deploys, and exploratory test-writing where natural language is easier than code. Also useful when debugging flaky tests — the skill captures artifacts and suggests root causes.
references/test-file-format.md and example .test.txt files in the repo.atr run --behavior <file|dir> to execute tests; use --headless=false to debug, --viewport to emulate devices, and --cdp-endpoint to connect to an existing browser. Best practices for element selection and waits are documented in the SKILL.md.Best suited for agents with browser automation capabilities (ATR, Claude Code or any agent that can drive Chromium via CDP). It assumes the environment can launch or connect to a browser and has filesystem access for test files.
This skill has not been reviewed by our automated audit pipeline yet.