second-brain-import-md

Second Brain — Add Markdown

Accepts pasted markdown content directly and writes it as a raw source file into raw/, ready for ingestion. Use this when you have content from any source (a web article, a copied document, notes, a transcript, an email) and want to get it into the knowledge base without exporting a file first.

Invocation

/second-brain-import-md [markdown content]
/second-brain-import-md "Optional Title" [markdown content]

If invoked with no content at all, ask: "Please paste the markdown content you'd like to add."

Execution

Step 1 — Parse arguments

Check whether the first token of the argument is a quoted string (starts and ends with "). If so, treat it as a title override and the remainder as the content body.

Otherwise, treat the entire argument as the content body and derive the title automatically (Step 3).

Step 2 — Validate content

If the content body is empty or only whitespace, report:

Error: No content provided. Please paste the markdown you'd like to add.

and stop. Write nothing.

Step 3 — Derive title and slug

If a title override was provided: use it directly.

Otherwise, derive the title from the content:

1. Look for the first # Heading in the content — use that as the title if found.
2. If no heading, use the first non-empty sentence or line (truncated to 60 chars).
3. If content is very short or unstructured, use "Untitled Note".

Generate the slug from the title:

• Lowercase all words
• Replace spaces and special characters with hyphens
• Remove characters that are not alphanumeric or hyphens
• Truncate to 50 characters

Step 3b — Detect content date

Scan the content body for signals about when the content was originally created or published. Look for:

• Explicit date stamps: PUBLISHED JAN 15 2026, Date: 2026-01-15, Updated: May 2026, March 2026, etc.
• Datelines in headings or sub-headings: # Q1 2026 Review, ## June 2026 MBR
• Inline temporal markers near the top of the document: "Published", "Written", "Last updated"
• Document title containing a year/month/quarter

If a date is found, convert it to YYYY-MM-DD format (use the 1st of the month when only month/year is available, e.g. May 2026 → 2026-05-01). Set content_date to this value.

If no date is found, leave content_date blank — do not guess.

Step 4 — Determine output path

raw/YYYY-MM-DD_<slug>.md

Where YYYY-MM-DD is today's date.

If a file at that exact path already exists, append a counter: raw/YYYY-MM-DD_<slug>-2.md, etc.

Step 5 — Write the file

Write the file with a minimal front-matter header followed by the content as-is:

---
source: pasted
imported: YYYY-MM-DD
title: <Derived or provided title>
content_date: YYYY-MM-DD        # omit this line entirely if no date was detected
---

<content body>

• imported is always today's date (when the file was added to the vault)
• content_date is the detected original creation/publication date of the content — omit the field if not found
• Do not modify or reformat the content body. Preserve all original markdown — headings, lists, tables, code blocks, links.

Step 6 — Confirm

Report:

✓ Markdown added

Title:         <title>
Output:        raw/<filename>
Size:          ~N words
Content date:  <YYYY-MM-DD if detected, otherwise "not detected">

Next step: run /second-brain-ingest to incorporate into the wiki

Invariants

• Only writes to raw/ (never raw/craft/, raw/pdf/, wiki/, or outputs/)
• Never modifies existing files — always creates a new file
• Never reformats or summarises the content — stores it verbatim

Error Conditions