Autonomous livecoding mode uses a priority-ordered if-then table where the first matching row fires

When an agent acts autonomously on a live set it runs a priority-ordered decision table each cycle and the first matching condition wins. High-priority rows handle failure and restraint first: recover a silenced/broken output before anything else, then wait after any recent save, then establish the intro before adding. Lower rows drive the arc: build or subtract one voice toward the target energy, fire a transition only at a section boundary, drop a self-clearing fill one bar before a boundary, add a micro-variation only when the groove has plateaued, avoid recently retired ideas, and wind down in the outro. The final default row, when nothing above fires, is to do nothing. Because the table is read top-to-bottom and the first match wins, every action is guarded and most guards enforce waiting, which structurally prevents over-editing.

Examples

If the groove is stable, on-arc, and no higher row applies, the correct autonomous action is the default row: lay out and do not save. A silenced output is handled by the top recovery row before any new idea is added.

Assessment

Given a scenario (e.g. groove good, energy on-arc, no build/subtract/boundary condition met), state the correct autonomous action and explain why the top-to-bottom, first-match structure produces it.