p2a-intakelisted

# Plan2Agent Intake Convert an early idea into structured planning input. ## Inputs - One-sentence product or feature idea. - Optional user notes. - Optional prior `intake_json` and newly answered decision ids when resuming. - Optional same-project and conditional cross-project Memory recall reports prepared by the harness for an iterative project. ## Output Return an `intake_json` object conforming to `.plan2agent/schemas/intake.schema.json` with: - `schema_version`: `p2a.intake.v1` - `idea`: original idea - `summary`: one paragraph restating the idea - `known_facts`: facts stated by the user - `assumptions`: objects with `id`, `statement`, `risk`, and `confirmation_needed` - `clarifying_questions`: objects with `id`, `question`, `why_it_matters`, and `blocks` - `needs_user_decision`: objects with `id`, `question`, `options`, `impact`, `default`, `status`, and optional `answer` - `evidence`: source objects with `source_id`, `title`, `url`, and `used_for` - `status`: `blocked_on_user` when any decision is `open` or `deferred`, otherwise `ready_for_spec` - Also produce a human-readable analysis in the conversation. The harness may generate `intake.md` as an optional view/export from `intake_json`, but `intake_json` is the source of truth. The analysis should follow the harness soft template and contain the restated understanding, each assumption with its reasoning, and for every `needs_user_decision` the option trade-offs, a recommended option with rationale, the downst