Reliable handoffs combine explicit design requirements, component mappings and rendered verification. Executable contracts can reduce repetitive checking for covered requirements; review still depends on model capability, integration quality and untested behavior.
An architecture for evaluating autonomous coding loops: durable state, bounded workers, independent acceptance evidence and honest human handoffs.
There are two ways to use a coding agent: sit in its terminal, or take its face off and drive the engine from your own code. If you're building anything around agents, the second mode is the whole game — and it's full of traps nobody warns you about.
Ralph-style loops launch fresh workers and keep state in files. How they differ from an in-session Stop-hook plugin, and what verification and isolation they require.
Native /goal, /loop and /schedule reduce orchestration work. Independent acceptance checks, scoped permissions and resource limits still determine reliable autonomy.
Claude Code cloud routines run scheduled work independently of your laptop. How to scope repositories, credentials and verification for durable automation.
Claude Code’s /loop schedules prompts inside a session. How fixed and dynamic schedules differ, what they retain, how to stop them and how to measure cost.
Claude Code’s /goal uses a separate evaluator of conversation evidence to guide continuation. The evaluator cannot independently run checks, and documented stopping and error conditions still apply.
Autonomy depends on the control system around an agent: scoped authority, durable state, bounded continuation and independent acceptance evidence. An illustrative design.
Claude Tag brings shared AI work into Slack. Understand channel identity, public and private memory, access controls and a practical rollout.
Ultracode combines xhigh reasoning with workflow orchestration. Current Claude Code supports a persistent ultracode setting; effort and permissions remain separate.
The June 2026 Fable 5 suspension exposed an engineering dependency: hosted-model access can change abruptly. What happened, what remains unproven, and how to design continuity.