compose-brownfield-prdlisted

# Compose Brownfield PRD Maps `characterize-brownfield` output into a PRD matching the exact HVE format so downstream HVE agents (GitHub Backlog Manager, `ado-prd-to-wit`, `jira-prd-to-wit`) can consume it directly. Loaded internally by `brownfield-analyst` after characterization completes — not designed for standalone dispatch (hence `disable-model-invocation`), but callable directly if characterization artifacts already exist from a prior run. **Boundary.** Consumes characterization artifacts; never re-scans the repo itself (reload the persisted artifacts — B4 — rather than re-deriving). Never creates issues. Never edits code. ## Inputs - Path to characterization artifacts (`index.md` + siblings from `characterize-brownfield`). - Product name (used for the kebab-case filename). - Goals (optional — if the human states explicit modernization goals, fold them into Section 1; otherwise derive draft goals from the "Modernization Opportunities" signal in tech-debt.md). ## Mapping table (characterization -> HVE PRD section) | Characterization source | HVE section | |---|---| | structure.md, index.md summary | S5 Product Overview, S1 context | | features.md (Core/Secondary) | S6 Functional Requirements (`FR-NNN`) | | features.md (Legacy-unused) + coverage.md gaps | S2 Problem (current situation), S4 Constraints | | integration.md | S9 Dependencies, S7 NFR (reliability/performance), S11 Privacy/Security | | contracts.md (if not `none-found`) | S6/S7 (API-shaped FR/NFR), S9 D