Verified Task Completion acts as a critical quality gate for AI-assisted development, ensuring that no claims of success are made without fresh, verifiable evidence. It establishes an 'Iron Law' that mandates running relevant commands—such as test suites, linters, or build scripts—and analyzing the full output before reporting status or committing changes. By preventing premature satisfaction and 'should-work' assumptions, this skill eliminates common failures like shipping undefined functions, missing edge cases, or relying on partial verification, ultimately maintaining high standards of software reliability and developer trust.
주요 기능
01Line-by-line requirements checklist verification
02Mandatory gate function before claiming success or satisfaction
03Real-time command execution and output analysis requirement
047 GitHub stars
05TDD Red-Green cycle verification for regression testing
06Independent validation of agent-delegated task results
