Results for “durable-execution”

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

Produces a machine-validated, auditable JSON decision record with assumptions, risks, recommendation, and review gating for high-stakes decisions.