forge-product-maplisted

Capture and maintain a project's product map — user journeys, scenarios, business rules, and workflow diagrams as curated Mermaid knowledge entries. Use this skill whenever someone wants to: document product flows, capture user journeys, generate scenario diagrams, maintain business rules visually, update the knowledge workspace, map user-facing routes, or bootstrap product documentation. Triggers on: /forge-product-map, capture product map, maintain user journeys, generate scenario flowcharts, document business rules, map product flows. IMPORTANT: This skill creates PRODUCT diagrams only — user journeys, UI routes, acceptance-criteria flows. It does NOT create architecture or codebase diagrams (no file paths, function names, or implementation details).
SidCorp-co/forge · ★ 4 · AI & Automation · score 73

Install: claude install-skill SidCorp-co/forge

# forge-product-map Maintains a project's curated product knowledge as Mermaid visual entries. Writes `overview`, `scenario`, `workflow`, and `rule` entries via `forge_knowledge action=upsert`. All diagrams represent **user-facing product behaviour** — never internal implementation. ## Modes - **bootstrap** (default) — full sweep: read existing entries, read shipped issues for evidence, identify gaps, ask the human only for gaps that cannot be inferred. Write/refresh all entry kinds. - **refresh** — incremental: read recently closed/released issues since the last run, update existing entries that those issues affect, add missing scenarios. - **gap [scenarioName]** — produce or fill exactly ONE scenario entry. Ask the human for the scenario name if not provided. ## Hard verification gate (AC5 — non-negotiable) Before writing ANY node into a scenario or workflow diagram: 1. Every node must map to one of: a real issue id, an acceptance criterion phrase, or a user-facing route string (e.g. `/projects/:slug/library`). 2. **NO** `file:line`, function names, module names, or source-code identifiers as nodes. If a node names a file or function, DELETE it — replace with the user-facing action it implements. 3. Tag each entry: - `confidence: "verified"` when a shipped (closed/released) issue directly backs the node. - `confidence: "inferred"` when derived from comments, ACs, or reasonable inference. 4. Store backing issue ids in `metadata.relatedIssueIds: string[]`. Violat