bootstrap-content-agentlisted

# Bootstrap a content agent You are turning this kit into a working, project-specific content agent. Work in phases. Do not skip the interview — the answers drive everything you generate. ## Phase 1 — Read the method Read `docs/01-architecture.md` and `docs/03-conventions.md` in full. Then skim `docs/02-playbook-spec.md`, `docs/05-publishing.md`, `docs/06-scheduling.md`, `docs/07-review-gate.md`, `docs/08-audit-maintain.md`, `docs/10-crawl-discovery.md` and `docs/11-social-posting.md`. If the agent will make video, also read `docs/20-video-backends.md`; if it will use skills or connectors from elsewhere, `docs/17-skills-registry.md` and `docs/18-ads-and-marketing.md`. Internalize the six parts (playbook / knowledge / state / scripts / discovery / schedule) and the non-negotiable conventions (env-only, review-gate, idempotent, cleanup). ## Phase 2 — Interview the user Ask (batch related questions; propose sensible defaults): 1. **Domain & goal** — what content, for whom, published where? 2. **Item types & schema** — what kinds of items (article / long serialized / case / …), and the exact fields each needs. Get one concrete example per type. 3. **Cadence** — how many per day; rotate phases? (phase = day mod N). 4. **Publish target** — is there an ingest API? Base URL + endpoint paths + auth. If none, is it social-only? 5. **Discovery** — should it crawl sources for ideas (crawl4ai)? If yes, list a few source index pages + the article-link pattern. (Copyright: excer