strategic-compact

To prevent context loss during long AI sessions, you need manual context compaction at logical task intervals. This approach preserves crucial memory through task phases rather than relying on arbitrary auto-compaction.

The best way to manage AI context before a complex task transition is to use proactive suggestions that prompt you to manually compact memory. This ensures implementation details remain fresh before you start writing code.

Tool usage tracking triggers context compaction reminders by monitoring your tool calls and providing periodic prompts to use the `/compact` command. Reminders are activated when a configurable threshold is met or at regular intervals thereafter.

Yes, you can customize the threshold for context compaction suggestions based on tool usage. This allows you to tailor when periodic reminders to use the `/compact` command are activated during your workflow.

You should use manual context compaction instead of auto-compaction when moving between distinct phases like research and implementation. It prevents mid-task context loss by allowing you to choose the exact logical interval to preserve memory.