otto-project-knowledgelisted

# Otto project knowledge operations Use the daemon-owned project-knowledge tools as the only write path. The canonical store is Markdown under `.otto/knowledge`, but current truth and its provenance must be changed atomically by Otto rather than by editing files directly. ## Read - At task start, consult the automatically injected active-page catalog. Use `list_project_knowledge`, `read_project_knowledge`, or `read_project_knowledge_root` for pages relevant to the task before broad repository searches. - Use `query_project_knowledge` when the relevant page is not obvious. Normal reads return only confirmed active pages. Include inactive pages only for an explicit knowledge review. - Read compiled truth, evidence, tags, and the complete timeline. Treat `superseded` pages as history, not current guidance. - Inspect `.otto/KNOWLEDGE.md` and the Markdown page in Git when the user needs to see the source or review a diff. ## Write - Use `record_project_knowledge` for a new fact. Include one of `decision`, `constraint`, `requirement`, or `architecture`, a precise title, rich Markdown compiled truth, evidence, and useful tags. Use a readable kebab-case id when a specific human slug matters. New pages start as `proposed`. - Use `record_project_charter` for a durable initiative. Include rich Markdown outcomes, scope, constraints, and acceptance criteria. Review status says whether the charter is trusted; delivery status independently says whether it is charter