关于
The Strict Task Verification skill eliminates assumptions and 'should work' claims by requiring fresh, empirical evidence for every assertion of success. It mandates the execution of specific verification commands—such as test suites, build scripts, or linter checks—before allowing a task to be closed or a commit to be made. By utilizing a specialized test-runner agent to handle verbose outputs, it prevents context pollution while ensuring that every success criterion is explicitly met and documented with actual command results.
