- pr Switch to Fable for long unattended runs
A task will run 30+ minutes largely unattended (an overnight migration, a big multi-file refactor, a workflow fleet).
- pr Let AskUserQuestion stall rather than guess on overnight jobs
You're running an overnight or unattended job that might hit an AskUserQuestion.
- pr Use auto mode for unattended runs, bypass only for sandboxes
You're starting a long or unattended run and don't want to approve 200 prompts.
- pr Clear context after a second failed correction
Claude has tried to fix the same thing twice and is still wrong.
- pr Cap dynamic workflow size before an uncertain ultracode run
You're about to make an ultracode / multi-agent request whose scale you're unsure about.
- pr Delegate routine coding to a lower-power subagent
A multi-file task has a lot of mechanical coding (boilerplate, edits, wiring) around a few real decisions.
- pr Use Artifacts instead of pasting long output
You have a long PR walkthrough, investigation, or dashboard-like result you'd otherwise paste into Slack or Notes.
- pr Split the writer from the reviewer
A session just built a feature or made a nontrivial change.
- W28 The week Claude Code learned to be left alone
A run of releases (2.1.200–2.1.203) all pull in one direction — safer defaults, real cost dials, and background work you can finally trust. Plus Artifacts opens to Pro and Max.
- 07-04 Fable 5: the ceiling, not the default
Claude Fable 5 is Anthropic's most capable model — and its most expensive. A researched guide to where it wins, when its premium is worth it over Opus 4.8, how to run it overnight, and how to prompt it — grounded in Anthropic's own docs and engineering posts.