Results for “seedance”

Orchestrate token-heavy research, coding, and testing by delegating bounded tasks to cheaper subagents while reserving Claude Fable for architecture, synthesis, and final review.

Tracks bidirectional confidence scores across five domains (TECH, OPS, JUDGMENT, COMMS, ORCH) for agents and users, with triggers, anti-patterns, and pool scoring per runtime.