| name | catalog-insights |
| model | haiku |
| description | Catalog book video knowledge into knowledge vault |
| disable-model-invocation | false |
| argument-hint | <book-slug> |
Catalog Insights
Extract knowledge from a completed book video and persist it in the Knowledge Vault for cross-book intelligence.
Context
- When to use: After video production is complete (voice, images, render done). Before publish.
- Input:
books/<slug>/notes.md + books/<slug>/storyboard.md
- Output: Updates to
knowledge-base/ files (append-only) + source added to Master notebook
- Pipeline position: After
/write-metadata, before publish. Also integrated into /produce-video Phase 6.5.
Pipeline Position
/produce-video or manual pipeline
↓
/write-metadata → metadata.md
↓
/catalog-insights → knowledge-base/ updates ← YOU ARE HERE
↓
Publish (Manual)
Steps
-
Validate: Check that $ARGUMENTS is provided. If missing, ask Hai for the book slug. Set SLUG=$ARGUMENTS. Verify projects/ai-book-video/books/$SLUG/notes.md exists.
-
Read inputs:
- Read
projects/ai-book-video/books/$SLUG/notes.md — key insights, author, angle, criticism
- Read
projects/ai-book-video/books/$SLUG/storyboard.md — scene structure, template used, emotional arc
- Extract: book title, author name, chosen angle, key concepts, blind spots, cross-references
-
Check duplicates: Read projects/ai-book-video/knowledge-base/library.md. If this book slug already exists in the table, warn Hai and ask whether to update existing entries or skip.
-
Extract concepts: For each key insight/concept from notes.md:
- List existing theme files:
Glob for knowledge-base/concepts/*.md
- Determine which theme each concept fits (habits, identity, productivity, focus, etc.)
- If concept fits existing theme: append a new
### entry to that file
- If concept needs a new theme: propose theme name to Hai via
AskUserQuestion, then create new file
- Each entry follows the format:
### [Concept name]
- **Source**: [Book Title] ([Author])
- **Position**: [Author's claim/argument]
- **Nuance**: [Criticism, limitation, or deeper insight from Bookie research]
- **Cross-ref**: → connections/[type].md #[anchor] (if applicable)
-
Extract author profile:
- Check if
knowledge-base/authors/<author-slug>.md exists
- If exists: append new book's data under a new
## Books Covered entry + update blind spots
- If not: create author file with: Background, Core Thesis, Strengths, Blind Spots, Books Covered, Cross-references, NotebookLM info
-
Identify connections: Read ALL existing knowledge-base/concepts/ files. For entries from OTHER books on overlapping themes:
- If positions contradict: append entry to
knowledge-base/connections/contradictions.md
- If positions agree/reinforce: append to
knowledge-base/connections/agreements.md
- If one extends/evolves another: append to
knowledge-base/connections/evolutions.md
- Each connection entry includes: Position A, Position B, Tension/Common ground, Discovered in
-
Record production metadata: Append row to knowledge-base/production.md:
| [Book] | [Template] | [Duration] | [Scenes] | [Hook line] | [Date] |
-
Update library index: Append row to knowledge-base/library.md:
| [Book] | [Author] | [Angle] | [Template] | [Date] | [Slug] |
-
Add to Master notebook:
- Find "Bookie: Library" notebook via
notebook_list
- Add
books/$SLUG/notes.md as text source via source_add
- If NotebookLM fails: log warning but continue — local files are source of truth
-
Report: Summarize what was cataloged:
- Concepts added (which themes, how many entries)
- Author profile created/updated
- New connections discovered (this is the exciting part!)
- Production metadata recorded
- Master notebook updated (yes/no)
- Suggest: "Next time you research a book on [overlapping theme], the vault will surface these connections automatically."
Important
- All concept/connection content in Vietnamese (matching notes.md language)
- Append-only operations — never delete or overwrite existing vault entries
- Cross-references use
→ arrow syntax with file path + anchor
- Theme files are indexed by THEME, not by book — one book may contribute to multiple themes
- When in doubt about theme classification, ask Hai
- The Knowledge Vault is the long-term competitive moat — quality of entries matters more than speed