What problem does it solve? Setting up a local environment for tests and QA is slow, inconsistent across machines, and repeated every time an agent or developer needs a running app. This Skill compiles one-time discovery of how a project runs into committed launch scripts, so every subsequent run boots the same environment in seconds on macOS, Linux, WSL2, or Windows. ## Core Features & Use Cases - Entrypoint script generation: Discovers the repo's own run recipe (services, build chain, ports) and generates test-env-up.sh/test-env-down.sh (or PowerShell .ps1 flavors) with locking, reuse checks, build caching, and health waits baked in. - Browser provider provisioning: Ensures the configured browser provider (e.g. Playwright or agent-browser) is installed and verified through its .ai/browsers/<provider>.md descriptor during generation. - Shared environment descriptor: Writes .ai/qa/test-env.json with base URL, services, credential references, and platform so QA and integration-test skills attach to the same running instance. - Self-improving repair loop: Any failure or degradation is patched back into the scripts, proven by re-running them, and logged in a # history: header. - Use Case: A coding agent needs a running app to execute UI tests. On first run the Skill discovers the stack, generates and verifies the up/down scripts cold and warm; every later run just executes the script and reports the base URL. ## Quick Start Ask the agent to prepare the test environment for this repository so QA and integration-test skills have a running app to attach to.
