What problem does it solve? Development work drifts out of sync with documentation: decisions lose their rationale, guides describe stale behavior, and autonomous agents resume work without context. This Skill treats docs/ as persistent AI context, keeping specs, guides, workstreams, issues, and ADRs aligned with the code as it changes. ## Core Features & Use Cases - Layered documentation governance: Routes every fact to exactly one layer — specs for intent, guides for current behavior, ADRs for decision rationale, workstreams/issues for active work — with templates and a validator enforcing the contracts. - Human-bounded autonomous work: Creates workstreams only after an interview confirms the authorization envelope, merge policy, runnability, and acceptance verification, so autonomous runs stop only at gates a human actually set. - Deterministic tooling: Scripts create issues, workstreams, and ADRs, archive completed work transactionally with index updates and rollback, and validate the whole docs tree including relative link resolution. - Use Case: When starting a feature, the agent reads docs/00_index.md, interviews you for scope and gates, creates a workstream, records design decisions as ADRs during implementation, updates the affected guide in the same slice, and archives everything after validation passes. ## Quick Start Ask the agent to create a workstream for your next feature so it interviews you about scope and gates, then keeps specs, guides, and ADRs updated as the code changes.
