awesome-second-brain: Command for Claude Code

Capture the following input.

1. Determine capture type from content:

   • Braindump: Free-form thoughts, ideas, observations
   • URL: If content contains a URL, use defuddle parse <url> --md to extract clean markdown
   • Document reference: If a file path (absolute or relative) or specific location is provided, read the file, extract its contents, and summarize into the capture note.

2. Determine destination — ask the user (or infer from content if obvious):

   • Inbox (default): ideas, facts, URLs, braindumps to be classified later → inbox/raw/
   • Thinking: reasoning, scratchpads, open questions, ongoing research → thinking/

3. Create the capture note with status: unprocessed (inbox) or status: thinking (thinking):

   Inbox (inbox/raw/[topic]-[YYYY-MM-DD].md):

   ---
   date: "[today]"
   created: "[today]"
   status: unprocessed
   ---
   

   Thinking (thinking/[YYYY-MM-DD-topic].md):

   ---
   date: "[today]"
   created: "[today]"
   status: thinking
   tags:
     - thinking
   ---
   

4. Structure content with observation categories where natural:

   • [fact] — Objective, verifiable information
   • [idea] — Subjective concepts, hypotheses
   • [decision] — Commitments, choices made
   • [technique] — Methods, processes, tactics
   • [requirement] — Constraints, dependencies
   • [question] — Open inquiries to investigate
   • [insight] — Realizations, pattern recognition
   • [problem] — Issues, pain points
   • [solution] — Fixes, workarounds
   • [action] — Task items

   Syntax: - [category] content

   Don't force categories — if the input is simple or short, plain text is fine.

5. Report what was captured and where it was saved.

6. Remind: For inbox captures, run /process to add frontmatter and classify, then /distribute to promote to a page. For thinking captures, use /process when ready to promote.

Content to capture: $ARGUMENTS