From claudian
Process the user's freeform notes in the vault's ideas/ folder into structured knowledge notes. Use when SessionStart flags unprocessed ideas, or when the user asks you to look at their ideas.
How this skill is triggered — by the user, by Claude, or both
Slash command
/claudian:vault-extractThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Extract structured knowledge notes from freeform ideas in the vault's `ideas/` folder.
Extract structured knowledge notes from freeform ideas in the vault's ideas/ folder.
ideas/ belongs entirely to the user — their private scratchpad. Claude reads it but NEVER writes into it. The only write allowed is setting processed: true in frontmatter after explicit user approval.
ideas/ without processed: true1. Read the idea. Understand what the user was capturing.
2. Identify extractable knowledge. One idea may yield multiple notes:
architecture notepattern notegotcha noteknowledge notePurely speculative or to-do ideas don't need extraction — just flag them processed with a note explaining why.
3. Propose structured notes. Show the user what would be created before creating anything:
Proposed note: <title>
Type: <type>
Folder: <destination folder>
Tags: [<tags>]
Summary: <1-2 sentence description>
List all proposals together before asking for approval.
4. Wait for approval. Always. Ask: "Ready to create these N notes. Should I go ahead, or would you like to adjust anything first?" Revise and re-show if changes are requested.
5. Create approved notes. Use vault-write for each. Set source: extracted in frontmatter. One call per note — never batch.
6. Flag the original idea. Update its frontmatter after all notes are created:
processed: true
extracted-to:
- "Title of First Note"
If skipped (no extractable knowledge):
processed: true
extracted-to: []
extraction-note: "Speculative — no durable knowledge to extract yet"
source: extracted. Never claude or human for notes extracted from ideas.Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub cyanotex/claudian --plugin claudian