Shows when to use planning, summarization and fresh-task handoffs to keep an AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition →'s context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition → on-task over long sessions. The patterns transfer to other coding agents.
Key takeaways · AI-distilled
When a request needs rewriting, Cline recommends editing the message that sent the task off course and choosing Restore Chat (keep workspace state) or Restore All (rewind files and chat), not arguing with a polluted thread.
/smol (or /compact) condenses history in place within the same task, a manual Auto Compact, while /newtask packages the plan, decisions, relevant files and next steps into a fresh task for a clean handoff.
Cline says the Focus Chain todo list persists through Auto Compact summarizations, so progress stays anchored after history is replaced; its reminder cadence defaults to every 6 messages and is configurable.
For multi-day or multi-person work, Cline recommends keeping a Memory Bank of Markdown files such as activeContext.md and progress.md current, then starting the next session with /newtask.
Terms in this piece · Glossary
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
