What problem does it solve? Personal archives scattered across Dropbox, Backblaze B2, Gmail takeouts, and old hard drives contain valuable writing, ideas, and correspondence that are never revisited. This Skill systematically explores those archives, filters out noise, and surfaces the content worth preserving in a structured knowledge base. ## Core Features & Use Cases - Safety-gated scanning: Refuses to run unless an explicit archive-crawler.scan_paths: allow-list is set in gbrain.yml, preventing accidental ingestion of sensitive files like tax documents or medical records. - Gold filtering and triage: Applies a keep/skip filter to separate personal writing, ideas, and relationship material from system files, binaries, and bulk mail, tracking every item's status in a manifest page. - Multi-format ingestion: Handles plain text, HTML, Markdown, .mbox email archives, .doc/.docx, .pst Outlook files, and compressed archives, filing ingested content into originals/, personal/, or ideas/ per the user's filing rules. - Use Case: Point the Skill at an old Dropbox archive of letters and journals; it maps the tree, proposes a priority queue, shows high-value items one at a time, captures your exact reactions, and creates cross-linked brain pages for each keeper. ## Quick Start Add the folders you want scanned to the archive-crawler.scan_paths: list in gbrain.yml, then ask the agent to crawl your archive and surface the writing worth keeping.
