Open Skills
For any system with an HTTP API, front-end UI, database, and third-party integrations: auto-analyze the architecture, generate a coverage checklist, and land tests across 12 ordered layers with multi-agent collaboration, independent acceptance, and hard pass gates.
$git clone https://gitee.com/interlingual-interaction/skill.git$cd skill/full-system-testingDetails
This skill enforces a strict set of rules for rigorous full-system testing: tests only report FAIL and never fill in missing business logic; each test type uses the appropriate industrial-grade tool rather than a hand-rolled equivalent; and all tests run against resettable seed data, restored to the initial state before every run.
The workflow starts with system analysis: answer 7 questions about page structure, APIs, database, and integrations first, use the answers to activate the matching test modules, then land tests across 12 ordered layers — with E2E execution, evidence collection, service management, and a structured event protocol.
The core principle is skeptical verification: don't trust an agent's self-report, don't trust the project's existing test files, and don't trust a pass claim without evidence. Missing key evidence can only be marked BLOCKED, never PASS.
The repository can hold multiple skills; this one lives in the full-system-testing/ subdirectory. Clone the repo and add that subdirectory to your AI IDE's skills folder to use it.