What problem does it solve? When an orchestrating AI (Architect) delegates implementation work to a worker AI (Executor), it is hard to know whether the delegation is actually paying off. This Skill provides a methodology for measuring delegation efficiency, spotting recurring defect patterns, and turning those findings into better task specifications. ## Core Features & Use Cases - ROI Analysis: Compare actual delegation cost against the hypothetical cost of generating the same work inline, including the Executor's own execution cost when the operator pays for its tokens. - Defect Tracking: Group defects found during verification by severity and category to identify recurring patterns such as missed edge cases or styling mismatches. - Cancellation Signals: Surface high cancellation or abort rates as a signal that a task spec needs tighter constraints or is a poor fit for delegation. - Feedback Loop: Propose concrete guardrails for future task specifications or targeted Executor-side skills to close recurring gaps. - Use Case: After a week of delegated coding tasks, ask the Architect to review the telemetry, report the savings multiplier versus inline generation, and recommend spec improvements for the task types with the most defects. ## Quick Start Analyze the delegation telemetry from my recent Executor runs and tell me whether delegation is paying off, which defect patterns keep recurring, and what guardrails to add to future task specs.
