Zettel Brainstormer 🧠

This skill formalizes the process of taking a rough idea or draft and enriching it with deep research, diverse perspectives, and structured brainstorming.

New Dual-Stage Workflow

This skill now supports a two-stage pipeline to balance cost and quality:

1. Preprocess & Extraction (cheap model)

• Run a lightweight extraction step using OpenRouter’s kimi-k2.5 to produce:
  • Short bullet-point keypoints
  • Candidate headings and tags
  • Minimal bibliography/search queries
• Output format (JSON): {"headlines":[], "points":[], "queries":[]}.

2. Draft & Humanize (pro model)

• Use a pro model (configurable; default: openai/gpt-5.2) to expand the best outline into a full draft.
• Post-process the draft with if exists, to make voice natural. Use the humanizer skill if available.
• Add proper Obsidian properties, tags, and links. Use the obsidian skill if available. Pick up properties and follow examples from other obsidian notes to keep the style consistent.

Files & Scripts

This skill includes the following resources under the skill folder:

• scripts/preprocess.py — call preprocess model, extract key points, write JSON
• scripts/draft.py — call pro model with outline + references, produce final Markdown
• scripts/wikilink_extractor.py — extract [[wikilinks]] from notes recursively
• references/templates.md — output templates, headline/lead examples, tone guide
• config/models.json — user-selectable model and research settings

Configuration & Setup

First Run Setup: Before using this skill, you must run the setup script to configure models and directories.

python scripts/setup.py

This will create config/models.json with your preferences. You can press ENTER to accept defaults.

Configuration Fields:

• pro_model: The model used for drafting (defaults to agent’s current model)
• preprocess_model: Cheap model for extraction (defaults to agent’s current model)
• zettel_dir: Path to your Zettelkasten notes
• output_dir: Path where drafts should be saved
• search_skill: Which search skill to use for web/X research (web_search, brave_search, or none)
• link_depth: How many levels deep to follow [[wikilinks]] (N levels, default: 2)
• max_links: Maximum total linked notes to include (M links, default: 10)

To change settings later, you can edit config/models.json directly or re-run scripts/setup.py.

Note: When using this skill, the agent will inform you of the current link_depth and max_links settings in the final result.

Usage

• Trigger when user asks: “brainstorm X”, “expand this draft”, “research and add notes to “.
• Example workflow (pseudo):
  1. If not configured, config will use agent’s current model by default (or run python scripts/setup.py for custom settings)
  2. Pick a random seed note from zettelkasten
  3. Extract wikilinks: scripts/wikilink_extractor.py --seed <seed_note> --depth N --max M --zettel-dir <zettel_dir> --output /tmp/links.txt
  4. Preprocess: scripts/preprocess.py --input <seed_note> --output /tmp/outline.json
  5. Research using configured search_skill (if not “none”)
  6. Draft: scripts/draft.py --outline /tmp/outline.json --model <pro_model> --out /tmp/draft.md
  7. (Optional) Humanize: humanizer --in /tmp/draft.md --out /tmp/draft.human.md
  8. Append result to seed note or save to output_dir
  9. Inform user: “Extracted wikilinks with depth=N, max_links=M using search_skill=”

Notes for maintainers

• Keep preprocess outputs small (200-600 tokens) to save cost.
• Ensure all external links are included in the References section with full titles and URLs.
• When appending, always include a timestamp and short provenance line (e.g., “Generated by writing-brainstormer; preprocessed with kimi-k2.5; draft with gpt-5.2”).

Example Output Format (unchanged)

... [Original Draft Content] ...

---
## 🧠 Brainstorming & Research (2026-02-07)

### 🧊 Core Metaphor: The Iceberg
[Expansion of the idea...]

### 📚 References
*   **James C. Scott:** *Seeing Like a State* — Discusses Metis vs Techne...
*   **Gall's Law:** [Link/Summary]

### 🐦 Signals
*   **@user on X:** Argues that [point] ... [Link]

Tags: #brainstorming #research #zettelkasten