What problem does it solve? Manually executing test cases and recording results is repetitive and error-prone, especially when mixing API tests, UI tests, and human-reported observations. This Skill acts as a dedicated test execution engineer: it takes your structured test cases, runs them step by step with real tools, and outputs a standardized execution record without inventing results. ## Core Features & Use Cases - Strict Step-by-Step Execution: Validates preconditions first (marking cases as blocked when unmet), then executes each numbered step without merging, skipping, or reordering. - Real Tool-Based Verification: Runs API cases via curl/Bash and UI cases via Playwright, including automated execution of agent-exec YAML blocks with placeholder injection, response capture, and assertion evaluation. - Four-State Judgment & Defect Logging: Classifies each case as passed, failed, blocked, or not executed, and records defects with phenomenon, deviation, severity level (critical to suggestion), and fix direction. - Standardized Reporting: Outputs a fixed 7-column Markdown table plus execution summary, and can save a full report file with defect, blocked, and unexecuted lists. - Use Case: Paste a batch of API test cases exported from Jira; the assistant runs each curl request, compares responses against expected assertions, and returns a complete execution table with pass rate and defect statistics. ## Quick Start Execute the test cases in my attached file and give me the execution record table with pass/fail status and defect details.
