remember

To turn user corrections into durable methodology notes, you capture friction signals explicitly or contextually, converting them into structured guidance stored under ops/methodology/ for cross-session learning.

The best way to capture friction signals is using contextual mode with an empty target, which mines recent conversation interactions to identify corrections and generate structured methodology notes.

Yes, you can scan stored sessions automatically using session mining mode, which detects recurring friction patterns and generates methodology notes with provenance and source mode for auditability.

You store methodology notes with provenance for auditability by writing structured outputs to ops/methodology/ with links to related notes and records of the source capture mode used.

No, it does not require explicit descriptions for every entry, as it supports contextual mode to mine recent conversations and session mining mode to scan past interactions automatically.

You should not use automated session mining when stored sessions are unavailable or when recent real-time conversational context contains the most relevant friction signals to capture explicitly.