execution-strategylisted

# Execution Strategy Given a unit of work and the surface it will execute on, name the method the executor runs and the report contract that method produces. A brief that names no method leaves its executor to improvise one. ## The five strategies | Strategy | Executor runs | |---|---| | `trivial` | the change directly; no test, because there is no runtime behavior to test | | `direct-tdd` | superpowers:test-driven-development inline: RED -> GREEN -> REFACTOR, failing test named before code | | `resume` | the superpowers chain entered at the supplied artifact: spec in hand -> writing-plans onward; plan in hand -> subagent-driven-development | | `superpowers` | the full chain: brainstorming -> spec -> writing-plans -> subagent-driven-development | | `delegate` | triages against this table, picks one of the other four, runs it | `delegate` never picks itself, must respect the surface table below, and names its choice in its report so the dispatcher knows which report contract applies. Excluded: superpowers:executing-plans -- it defers to subagent-driven-development wherever subagents are available, and dispatched workers are full Claude Code sessions, so it is dominated. ## Picking the strategy - `trivial` -- no runtime behavior to test: pure docs, comments, config, or a mechanical rename with zero logic change. - `direct-tdd` -- real code with clear criteria and an existing pattern. - `superpowers` -- everything else: new features, multiple valid approaches, vague c