tomashrdlicka/process-links

process-links

Process all queued links from the web capture queue.

Usage

/process-links
/process-links --limit 5
/process-links --topic ai-ml

Options

• --limit N - Process only the first N links
• --topic TOPIC - Only process links that would go to this topic
• --dry-run - Show what would be processed without doing it

Instructions

When the user invokes /process-links, follow these steps:

Step 0: Load Configuration

Read config.json from the engram project root to get the vault path:

{
  "vault_path": "~/Documents/Obsidian/WebCapture"
}

Use the vault_path value as {VAULT_PATH} in all paths below. Expand ~ to the user's home directory.

Step 1: Read Queue and Vault Context

Read both files:

• data/queue.json
• {VAULT_PATH}/_system/index.json (for connection discovery)

Step 2: Filter Links and Deduplicate

• Only process links with status: "pending"
• Deduplicate against vault: For each pending link, check if its URL (normalized: strip trailing slash, fragment, tracking params like utm_*, ref, fbclid) already exists in index.json notes. If a match is found, mark the queue item as status: "processed" with note_path pointing to the existing note and skipped_reason: "duplicate", then skip it. Report skipped duplicates in the summary.
• Apply --limit if specified
• Apply --topic filter after determining topic (if specified)

Step 3: Process Each Link

For each pending link:

1. Report: "Processing: {title} ({url})"

2. Use the same logic as /capture:

   • Fetch content with WebFetch
   • Determine type and topic
   • Generate summary, key points, and enriched tags (6-12 tags, analyzing content beyond user-provided tags)
   • Find related notes - search index for notes sharing 2+ tags, same topic, or overlapping themes. Select top 3.
   • Add bidirectional wiki-links - in the new note's ## Connections section, add wiki-links with prose context. Read each related note file and append a backlink.
   • Create note file with ## Connections section
   • Update index (including related array on new note AND on related notes' entries)
   • Update topic page ## Notes section

3. If user_context is present in the queue item, use it to guide summarization

   Intent detection (per-link):

   • If queue item has share_twitter: true, add tag share-twitter and set share_intent: "twitter" on the index entry
   • If queue item has deep_learn: true, add tag deep-learn and set deep_learn: true on the index entry
   • Auto-detection fallback: If user_context contains phrases like "post on X", "post on twitter", "share on twitter", "postable on X", "tweet this", "share this", set share_intent: "twitter". If it contains "learn", "study", "deep dive", "important", "reread", "want to learn", "understand this", set deep_learn: true

4. After successful processing:

   • Update the queue item's status to "processed"
   • Add processed_at timestamp
   • Add note_path with the created note path

5. If processing fails:

   • Update status to "failed"
   • Add error field with reason
   • Continue to next link

Step 4: Save Updated Queue

Write the updated queue back to data/queue.json

Step 5: Regenerate Views

After all links are processed (not per-link), regenerate the 4 view files once from index.json data:

1. views/by-date.md - List all notes sorted by capture date (newest first). Group by date. Each entry: - [{type}] [[{note_path}|{title}]] - {summary snippet}

2. views/by-type.md - Group notes by content type (article, video, x_post, photo, tool, research, quick). Under each heading, list notes sorted by date.

3. views/unread.md - List only notes where read: false, sorted by date. Each entry includes title, type, topic, and capture date.

4. views/favorites.md - List only notes where favorite: true, sorted by date.

5. views/twitter-queue.md - Notes with share_intent: "twitter" and twitter_posted != true. Each entry: title with source URL link, 1-2 sentence summary, suggested tweet angle based on key_points/user_context, tags for hashtag inspiration. Header shows count.

6. views/learnings.md - Notes with deep_learn: true, grouped by topic. Each topic group shows: topic name with note count, each note with title/summary/connections to other learning resources. Footer suggests reading order.

Read {VAULT_PATH}/_system/index.json, filter/sort the notes array, and write each view file. Use Obsidian wiki-links ([[path|title]]).

Step 5b: Git Sync Vault

After all links are processed and views regenerated, push vault changes to git:

cd {VAULT_PATH}
git add -A
git commit -m "vault: processed N links (YYYY-MM-DD)"
git push origin main

If the vault is not a git repo yet, skip this step silently.

Step 6: Report Summary

Tell the user:

Processed X links:
- [title1] -> content/topic/note.md (connected to: note-a, note-b)
- [title2] -> content/topic/note.md (connected to: note-c)

Failed: Y links
- [title3]: error reason

Remaining in queue: Z links

Dry Run Mode

If --dry-run is specified:

1. Fetch and analyze each link
2. Show what would be created (type, topic, enriched tags, likely connections)
3. Don't create files or update queue
4. Report the plan

Error Handling

• If queue file doesn't exist, create it with empty array
• If a link fails, log the error and continue with remaining links
• Report all failures at the end

Notes

• Queue file: data/queue.json
• Vault: {VAULT_PATH}/
• Use the capture skill logic for each link, including connection discovery
• Always read the vault CLAUDE.md at {VAULT_PATH}/CLAUDE.md for vault conventions