What problem does it solve? AI assistants and developers often claim work is done, fixed, or passing based on assumptions, stale test runs, or partial checks. This Skill eliminates unverified completion claims by requiring fresh command output as evidence before any status assertion, commit, or pull request. ## Core Features & Use Cases - Evidence-First Gate: Enforces a five-step gate (identify, run, read, verify, claim) before any completion statement. - Context-Specific Checklists: Defines required verifications after implementing a task, before opening a PR, and after a bugfix, including full test suite, linter, and build runs. - Rationalization Blocking: Lists red flags and common excuses ("should work", "linter passed") with their counterarguments to prevent premature claims. - Use Case: Before opening a PR, the assistant runs the full test suite, linter, and build, then reports results with concrete evidence such as "Tests: 47/47 pass, Linter: clean, Build: exit 0" instead of saying "done". ## Quick Start Ask the assistant to verify the fix with fresh test output before claiming the bug is resolved.
