[Manohar Paturi (ManoharPaturi)]

Motivation & Context

to_markdown() writes - none yet as the placeholder for an empty memory list, and from_markdown() filtered every bullet equal to that text so the placeholder would not become a memory. The filter could not tell the placeholder from a real memory whose text happened to be exactly none yet, so that memory was silently dropped on every read. Reproduced on main: a topic written with ["none yet", "real fact"] reads back as ["real fact"], and the loss repeats on every reparsing pass.

Description & Review Guide

• What are the major changes? Parsing now keeps every bullet, and the placeholder is removed once after the loop, only when it is the sole entry.
• What is the impact of these changes? That sole-entry shape is the only one the write side produces for an empty memory list, so empty lists still round trip to empty. A genuine memory equal to the placeholder text survives.
• What do you want reviewers to focus on? The two round trip tests: the mixed list keeps both memories (fails on main), and the empty list still reads back empty.

Related Issue

Fixes #9725

AI Assistance

• No material AI assistance was used.
• This is an AI-assisted contribution. I reviewed, understood, and verified all submitted content and accept responsibility for it.