plan-to-prdlisted

# Plan to PRD Synthesize an already-interrogated plan into a **PRD** — the committed, durable specification of WHAT an Initiative delivers, written to `docs/prds/<slug>.md`. This is a near-zero-interrogation transform: the heavy interrogation already happened upstream. Do not re-run forensic questioning here. Read the plan, draft the PRD, run the two confirm gates, write the file. A PRD describes **WHAT** the Initiative delivers — never **HOW**. File scope, step sequencing, and the implementation path are the Directive's job, not the PRD's. The PRD is path-agnostic: producing it does not decide single-branch vs brood and does not force any execution path. That routing is decided downstream at implementation time, independent of this artifact. --- ## Inputs Two inputs, resolved in this order: 1. **The plan (required).** Source it from EITHER: - the **live conversation context** (an interrogated plan already present in this session), OR - a persisted plan file at `.hivemind/plans/<slug>.md`. If neither is present, stop and ask the user for the plan or its slug. Do not synthesize a PRD from an un-interrogated idea. 2. **A handoff (optional).** If a handoff file exists at `.hivemind/handoffs/<slug>.md`, read it for volatile session state (locked decisions, open questions, pointers). Treat it as supplementary context that points at the plan — it does not replace the plan. ### Deriving the slug Derive `<slug>` from the input; never hardcode it. - If sourcing fro