What problem does it solve? Starting service design directly from a raw intent document risks jumping straight to designed services without validating whether the intent's claims hold up. This Skill reads a repo's intent document and produces a research agenda -- candidate services to test and claims to test -- where every item traces back to a specific intent field, so downstream discovery work validates hypotheses against real discourse instead of rubber-stamping assumptions. ## Core Features & Use Cases - Distill gate: Cold-reads the intent against distill checks and halts with specific gaps if the intent cannot be acted on from its own text, before any agenda is derived. - Agenda derivation: Produces services_to_test entries ({name, actor, goal, test}) and claims_to_test entries ({claim, who, evidence_target, right_if}), each traced to an intent field such as what_it_is, cut_or_broken, open_questions, or evidence. - Prediction instrumentation: Emits one staked prediction per claim via legion uncertainty emit, with confidence anchored to the intent's own evidence fields, so sd-discover can later witness each claim as supported, bounded, or contradicted. - Use Case: At the start of a repo's service design, invoke this Skill on the intent document to get a structured agenda of hypotheses with prediction ids, then hand that agenda to sd-discover for evidence-based validation. ## Quick Start Review the intent document for this repo and produce a research agenda of services and claims to test before any service design begins.
