← ClaudeAtlas

slipway-plan-auditlisted

Use when validating that the governed artifact bundle is ready for execution. Triggers on post-authoring audit or whenever plan artifacts change materially.
signalridge/slipway · ★ 13 · AI & Automation · score 66
Install: claude install-skill signalridge/slipway
# Plan Audit ``` IRON LAW: NO EXECUTION WITHOUT A VERIFIED, COMPLETE PLAN ``` ## Purpose Validate that the governed artifact bundle is ready for execution. This host is the execution-readiness gate before implementation. ## Workflow Outline Read the bundle and codebase-map context, audit artifacts/dimensions/task shape, then write verification and wait for approval. ## Read Context Run `slipway next --json`, locate the governed change bundle, and read the durable codebase map when present. If `input_context.codebase_map_dir` contains durable mapping documents, read at least: - `ARCHITECTURE.md` - `STRUCTURE.md` - `TESTING.md` - `CONCERNS.md` Use them to verify task targets, blast radius, and test scaffolding. If the map is absent, continue and record the missing brownfield context as an advisory gap. If scope remains ambiguous after reading the bundle, stop and return to intake clarification. Before auditing task targets against the map, check `input_context.codebase_map_status`. A `scaffold_only` or `baseline` status means the documents are **non-durable** (template placeholders or CLI-detected facts only); do not audit against them as if they were reviewed context. Record the consume-time codebase-map advisory that `next`/`run` emits as an advisory gap rather than treating the map as complete brownfield analysis. A `populated` or `partial` status reflects content presence, not scope relevance — a map authored for a prior change still reads `populated`. When `next`/`