--- name: onboard-repository description: Skill to onboard a repository to this harness --- # Onboard Repository Skill ## Process - Use `get-context`, then the applicable `context-*` skills in this order: requirements, architecture, security, user experience, quality, coding, and `document-context`. - An area applies when it has relevant requirements, decisions, risks, quality expectations, or user flows. Create the root index and an index plus repository-local template for each applicable area; mark each non-applicable area `N/A` with a reason. - Use indexes for scope and navigation. Group related records by topic; write records as readable statements of intended outcomes, with IDs and rationale or verification details only where useful. - Establish the initial normative context from confirmed stakeholder intent. Clarify blocking gaps; leave unresolved context incomplete rather than inventing intent. Keep implementation observations and details out of specifications and templates. - After onboarding, every non-context-neutral implementation change requires an update to its relevant context first. ## Verification - Use the repository's quality context and existing conventions to identify useful checks. Add repeatable scripts only when valuable; keep scripts outside `/context` and expectations inside it. - Review indexes, templates, IDs, references, completeness, and open questions manually. Validators support but do not replace this review. Onboarding is complete when the root index links to applicable areas; each applicable area has an index and template; `N/A` areas have reasons; records are grouped, uniquely identified, and traceable; open assumptions and blockers are visible; and verification results and limitations are reported. If a clarification blocks onboarding, report the blocker and continue with independent areas when possible.