What problem does it solve? Writing comprehensive test cases from requirements documents is slow and inconsistent, and cases written for humans often cannot be executed by AI agents. This Skill converts PRDs, API documentation, and UI designs into standardized test cases that humans can import into ZenTao/Tapd/Jira and AI agents can execute directly via curl or Playwright. ## Core Features & Use Cases - Dual-Mode Output: Generates Markdown test case sets by default, or a single runnable pytest .py file when Python trigger words (pytest, 转 py, 可执行脚本) are detected. - 9-Dimension Coverage: Systematically covers normal flow, exceptions, boundary values, null values, special characters, permissions, uniqueness, repeated operations, and malformed inputs, with a coverage matrix. - AI-Executable Blocks: Each case carries an agent-exec YAML block with concrete request/selector data and assertions, executable via curl, Bash, or Playwright. - Use Case: Paste a login API spec and login page design; receive TC-API and TC-UI case tables with P0-P3 priorities, plus a pytest file where automatable cases become test functions and manual cases become annotated SKIP-MANUAL comment blocks. ## Quick Start Ask the AI to generate test cases from your requirement document, for example: generate test cases from this login API documentation and output a pytest file.
