| name | godworld_batch_ingest |
| description | Safely fetch, deduplicate, and ingest an entire Google Drive folder of GodWorld canon files into Supermemory. |
GodWorld Batch Drive Ingestion
When the user asks to "batch ingest", "sweep", or "process" a Google Drive folder full of articles or canon files, use this skill to fetch, deduplicate, and ingest them automatically without creating duplicates.
Step 1: Run the Batch Ingestion Script
Execute the batch script, passing the Google Drive Folder ID as the first argument.
node /root/GodWorld/.agents/skills/godworld_batch_ingest/scripts/batchIngestDriveFolder.js <FOLDER_ID>
What the script does:
- Calls
listDriveFolder.js to get all files in the directory.
- Filters for
.txt and .md files (ignoring subdirectories, PDFs, etc.).
- Fetches each file locally using
fetchDriveFile.js.
- Checks
/v3/search in Supermemory using the first 150 characters of the file to see if an identical chunk is already ingested (Score > 0.80).
- If not a duplicate, runs
ingestEdition.js to ingest the file into Supermemory.
- Writes a detailed log to
changelog.md in the current working directory.
Step 2: Review Changelog
Read the generated changelog.md in your current working directory and summarize the results for the user (how many were skipped as duplicates, how many were newly ingested).
