Capture

Capture

Use this skill when the user says to remember, save, capture, or drop something into the brain.

Upstream GBrain uses gbrain capture as the front door. This Codex port maps that idea to the repo's file-backed helpers first, then uses the upstream CLI when it exists.

Workflow

1. Identify the content to capture:
   • inline user text
   • a local file
   • a short summary generated during the current task
2. Pick the right capture path:
   • use scripts/brain_capture_signal.py for short user phrasing or ideas
   • use scripts/brain_ingest_source.py for files and source material
   • use specialized skills for meetings, media, or article-style ingestion
3. Before creating a new brain page, use brain-taxonomist when the target folder or page type is not obvious.
4. Include source context:
   • who said it or where it came from
   • date
   • citation or local file path when available
5. Return a short receipt with the created or updated path.

Guardrails

• Do not store secrets.
• Do not silently rewrite existing brain pages.
• Do not use shell inline arguments for sensitive content.
• If the user asks for a bulk import, route to media-ingest, idea-ingest, or sync-gbrain instead.