memory-ingest

To ingest raw files into a wiki, place them in the memory_root/raw directory. The system classifies each file, archives originals to memory_root/sources by topic, and delegates wiki content creation to memory-add.

To avoid duplicate files when archiving to a knowledge management wiki, the system checks the memory wiki log before processing. This ensures files already ingested from memory_root/raw are not re-archived into memory_root/sources.

File classification for data archiving works by selecting an appropriate topic and subtopic for each file. Originals from the raw directory are then archived under the corresponding topic path in memory_root/sources.

You cannot process multiple files at once for wiki ingestion because the workflow handles one file at a time. Files are placed in memory_root/raw and processed sequentially to ensure accurate topic classification and archiving.

You do need bash commands to manage file I/O for memory ingestion. The workflow uses absolute paths with bash to move originals from raw to sources and coordinate updates with the memory-add skill.