概要
This skill acts as a metacognitive guardrail that prevents Claude from making premature success claims by requiring tangible proof of completion. It forces the AI to execute builds, run unit tests, and demonstrate actual output before concluding a task, ensuring that code changes are truly functional, bug fixes are verified against original issues, and factual claims are cross-referenced with external sources or project documentation. By implementing a mandatory verification checklist, it eliminates the 'it should work' syndrome and improves the reliability of AI-generated implementations.