## Notes Rewritten based on discussion in: #3586240 This issue could be changed to a discussion issue or a design/UX issue. ## Summary Provide tooling to analyze a large context item and suggest subcontext splits for editorial review. Splits should follow logical topics/intent, not headings alone. Output creates or links proposed child items under a parent; editor confirms before save. ## Problem/motivation Large context items are easy to create but hard to use well at injection time. When a single item covers a broad topic, agents often receive more text than the task needs, which increases token use and makes it harder for the model to focus on what matters. Subcontext hierarchy already lets teams attach focused child items to a parent, but splitting content is still a manual editorial task: someone has to decide where to break material apart, create parent/child items, and keep relationships coherent. That overhead discourages splitting even when it would help token management and relevance. A practical next step is tooling that analyzes an existing context item and proposes logical subcontext splits based on topic and intent—not just Markdown headings—so editors can review, adjust, and save a parent/child structure instead of doing all of that by hand. ## Solution TBD ## Tasks - Discuss UX and architecture - Create design/UX/prototype - Implement - Review and test ## AI usage - [x] AI assisted issue