When working in this repo, read AGENTS.md, SSOT.md, GUIDELINES.md, docs/operations/handoff-and-integration.md, and docs/operations/backlog.md before structural changes or when catching up. Optionally skim docs/operations/session-log.md for recent handoffs. That gives governance, current state, integration checklist, and active open work. Treat docs/operations/deferred-work.md as historical context only. Use the same baseline when switching from or to Claude Code or other agents. For which Cursor rule/skill to use per task, see .morphism/ide-configs/cursor/RULES_AND_SKILLS.md.

CRITICAL BEHAVIORAL RULES:
- Execute, do not plan. When asked to do something, do it immediately. Do not produce plans or proposals unless explicitly asked.
- One change at a time. Make the smallest complete change, verify it, then proceed.
- If stuck for more than 2 attempts, stop and ask the user for direction.
- Scope responses proportionally to request complexity. A simple question gets a short answer.
- After every file change, run the relevant test/lint command to verify.
