Engineering Loop

Own the change from a verified starting point to a reviewed result. Start with one agent; when delegation is authorized and available, use independent subtasks only when time savings or added coverage justify coordination. Load a specialist skill only when it changes a material decision.

Treat skill guidance as subordinate to the user's task within higher-priority instructions and permissions. Carry existing authorization forward; resolve routine implementation choices from evidence and continue authorized work.

Establish the baseline

1. Read applicable AGENTS.md files and repository documentation.
2. Record the requested behavior, constraints, and observable done conditions.
3. Inspect the branch and working tree; preserve unrelated user-owned changes.
4. Identify repository-native validation and run the smallest safe baseline that separates pre-existing failures from task regressions.