What problem does it solve? Software can pass every acceptance test and still fail the user, because the requirements specification itself may be wrong. This Skill judges working software against the original statement of need rather than its acceptance criteria, surfacing where the specification diverged from what users actually wanted. ## Core Features & Use Cases - Need-based evaluation: Recovers the originating statement of need and walks the working software as the user pursuing their real goal, not the specified steps. - Classified, evidenced findings: Each gap is typed (unmet need, wrong target, missing requirement, over-specification, or stale assumption), backed by cited evidence, and priced by change cost. - Bounded prioritized report: Produces five to ten findings with an explicit verdict — MEETS THE NEED or GAPS FOUND — without modifying any specification or code. - Use Case: After a release is complete and demonstrable, ask whether the software fulfills its goals; the agent walks the running increments and returns a report your team feeds into the next requirements refinement round. ## Quick Start Ask the agent: did we build the right thing, and what gaps can you find in the requirements specification?