memory

A persistent memory profile stores your product context, stack, stakeholders, and roadmap state in a user-profile.md file, ensuring the assistant retains institutional knowledge and personalizes responses across future sessions.

The best way to parse informal gossip is using a gossip mode feature that extracts stakeholder notes, blockers, and timeline shifts, converting them into structured memory entries before requiring explicit user confirmation to commit the writes.

Yes, you can save decisions and open questions directly to product memory after drafting a PRD by appending new dated entries to the user-profile.md file, ensuring historical context is preserved while updating the last_updated timestamp.

Updating a persistent PM profile does not delete outdated roadmap context; it enforces an append-first approach that resolves items rather than deleting them, preserving historical decisions while maintaining current product context.

Staleness checks detect aged items in the persistent memory file by applying staleness thresholds, surfacing at most two prompts per session to review outdated risks, and requiring explicit user confirmation before committing any profile updates.

Yes, you need a user-profile.md file to track product context and lessons learned, as the memory skill reads and writes structured PM data to this specific markdown file to maintain accurate and up-to-date institutional knowledge.