Files
My-Brain-Is-Full-Crew/docs/agents/scribe.md
gnekt b1c99bacc8 Extract 13 skills from agents and update full documentation
Architecture change: complex multi-step flows (onboarding, email triage,
transcription, etc.) are now skills that run in the main conversation
context instead of agent subprocesses. This fixes the state/context loss
that caused agents to skip phases during multi-turn conversations.

Skills created (13):
- Architect: /onboarding, /create-agent, /manage-agent, /defrag
- Postman: /email-triage, /meeting-prep, /weekly-agenda, /deadline-radar
- Transcriber: /transcribe
- Librarian: /vault-audit, /deep-clean, /tag-garden
- Sorter: /inbox-triage

Agent changes:
- architect.md: -70% (1554 → 473 lines)
- transcriber.md: -72% (530 → 147 lines)
- postman.md: -42%, librarian.md: -37%, sorter.md: -11%
- All agents: explicit post-it create-if-not-exists

Scripts:
- launchme.sh/updateme.sh: copy skills/ directly, remove generate-skills.py

Docs updated:
- README.md: new Skills section, mermaid diagrams, routing
- getting-started.md, examples.md: skill references
- docs/agents/*.md: capability tables with skill vs agent routing
- references/agents.md, agent-orchestration.md, agents-registry.md,
  agent-template.md: skill registry, skill-first routing protocol
2026-03-25 11:56:52 +01:00

6.2 KiB

Scribe

Your fast-thinking, clean-writing note-taking partner.

What it does

The Scribe turns messy, rushed, half-formed thoughts into clean Obsidian notes. You type fast, you skip punctuation, you mix languages, you abbreviate everything. The Scribe does not care. It takes whatever you throw at it and produces polished, well-structured notes ready for your vault.

It is smart about what it captures. If you dump three unrelated topics in one message, it splits them into three separate notes. If you paste speech-to-text output full of "um" and "like," it cleans it up while keeping your voice. If you are rapid-firing ideas in brainstorm mode, it numbers them all without judging any.

Every note the Scribe creates lands in your Inbox (00-Inbox/), properly formatted with YAML frontmatter, tags, wikilinks to people and projects, and suggested connections to other notes. The Sorter will file them later. The Scribe's job is pure capture speed and quality.

Capabilities

  • Standard capture: classifies your input (idea, task, note, person, reference, list) and applies the right template
  • Voice-to-note: detects speech-to-text output and cleans it up by removing filler words, restoring punctuation, and preserving your natural voice
  • Thread capture: when you send a stream of consciousness, splits it into individual atomic notes linked together
  • Quote capture: formats quotes with author, source, page number, and your commentary
  • Reading notes: structures book/article/podcast notes with key takeaways, section-by-section summaries, and action items
  • Brainstorm mode: captures everything without filtering, numbers ideas, groups by theme, highlights the most promising ones
  • Multi-note detection: automatically splits unrelated topics into separate notes
  • Emotion tagging: detects emotional tone and tags notes accordingly
  • Auto-suggested connections: mentions 2-3 related notes or topics at the end of each note
  • Code, math, and diagram support: wraps code in fenced blocks, uses LaTeX for math, creates Mermaid diagrams
  • Multilingual: works in any language, handles mixed-language input gracefully

How to use it

Just start typing what you want to capture. The Scribe detects the right mode automatically. You can also be explicit:

  • "Save this: [your text]"
  • "Quick note: [your text]"
  • "Brainstorm: [topic]"
  • "Quote: [quote text]"
  • "Reading notes for [book/article title]"
  • "Thread: [stream of thoughts]"
  • "Voice note: [pasted speech-to-text]"

In Italian: "Salvami questo", "Nota veloce", "Appuntati" In French: "Sauvegarde ca", "Note rapide" In Spanish: "Guarda esto", "Nota rapida" In German: "Schnelle Notiz", "Schreib das"

Example conversations

Quick messy capture

You: save this: meeting with Marco, budget discussion, he wants report by friday, also mentioned new hire starting in april, seems worried about Q2 numbers

Scribe: I identified 2 distinct topics and created 2 separate notes:

1. Meeting With Marco, Budget Discussion Filed as 2026-03-21 - Note - Budget Discussion With Marco.md in 00-Inbox/

Key points captured:

  • Marco wants the budget report by Friday
  • He expressed concern about Q2 numbers
  • Action item: deliver report by Friday

2. New Hire Starting in April Filed as 2026-03-21 - Note - New Hire April.md in 00-Inbox/

Linked to 05-People/Marco in both notes. Does this capture everything correctly?


Brainstorm mode

You: Brainstorm: ways to improve the onboarding experience for new users

Scribe: Created 2026-03-21 - Brainstorm - Onboarding Improvements.md in 00-Inbox/

Captured all your ideas, numbered and grouped. I noticed 3 natural clusters: UI changes, documentation improvements, and community-driven ideas. The two most promising ideas (based on feasibility and impact) are highlighted in the Hot Takes section.

Want to expand on any of these?


Voice-to-note

You: Voice note: so basically i was thinking about um the whole pricing structure right and like marco said yesterday that we should probably move to a tiered model because uh the enterprise clients are paying the same as the small ones and thats just not sustainable you know so maybe we do three tiers basic pro and enterprise and each one gets different levels of support

Scribe: Detected speech-to-text input. Cleaned and structured as:

Pricing Structure, Tiered Model Proposal

Marco suggested moving to a tiered pricing model because enterprise clients currently pay the same as small clients, which is not sustainable. The proposed tiers:

  • Basic: entry level
  • Pro: mid-tier
  • Enterprise: full support

Each tier would include different levels of support.

Filed as 2026-03-21 - Idea - Tiered Pricing Model.md with source: voice-note in the frontmatter. Sound right?

Works with

  • Sorter files the notes the Scribe drops in the Inbox
  • Architect provides the templates and naming conventions the Scribe follows. If a new project or area seems to be emerging, the Scribe flags it for the Architect
  • Connector gets flagged when the Scribe notices a note relates to many existing notes, so it can do deep linking

Tips

  • Do not overthink your input. The messier, the better. That is literally what the Scribe is for. Just dump your thoughts.
  • Use "quick dump" for multiple topics. The Scribe will detect and split them automatically.
  • Paste raw speech-to-text directly. No need to clean it up first. The Scribe handles filler words, broken sentences, and transcription artifacts.
  • Say "brainstorm" when you want zero filtering. The Scribe will capture every idea without judging, then gently highlight the best ones.
  • Check the suggested connections. The Scribe adds a Connections section at the end of each note. These are quick wins for building your knowledge graph.
  • Trust the assumptions. The Scribe makes judgment calls when your input is ambiguous. It always tells you what it assumed, so you can correct it if needed.

What it remembers

The Scribe keeps a post-it in Meta/states/scribe.md with notes from its last run: which notes it created, brainstorm topics in progress, and assumptions it made. This means if you come back to continue a brainstorm or refine a note, the Scribe already has context.