cold-start

Imports contacts, calendar, email, and archives into a new GBrain knowledge base.

Updated Aug 17, 2026
One-click install
npx skills add https://github.com/AMC-JTC/gbrain-1 --skill cold-start-amc-jtc
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cold-start
Source: https://github.com/AMC-JTC/gbrain-1/tree/main/plugin/skills/cold-start
Command: npx skills add https://github.com/AMC-JTC/gbrain-1 --skill cold-start-amc-jtc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? A freshly installed GBrain knowledge base is empty and useless until it contains your real data. This Skill sequences the highest-leverage data sources so a new brain goes from zero pages to a useful personal knowledge graph in one guided session. ## Core Features & Use Cases - Prioritized Import Pipeline: Walks through eight ranked phases covering markdown/Obsidian vaults, Google Contacts, Calendar, Gmail, conversation exports, X/Twitter archives, file archives, and meeting transcripts. - Credential Safety via ClawVisor: Routes all Google API access through a credential gateway so the agent never holds raw OAuth tokens, with offline Google Takeout exports as the fallback path. - Entity Cross-Linking and Resume: Detects people and companies during every import, links them into the knowledge graph, and tracks progress in a state file so interrupted sessions resume cleanly. - Use Case: After running gbrain setup, ask your agent to bootstrap the brain; it imports your Obsidian vault, seeds people pages from contacts, builds a meeting history from calendar events, and verifies retrieval with test queries. ## Quick Start Ask your agent to cold start the brain by importing my notes, contacts, calendar, and email after finishing gbrain setup.

Frequently Asked Questions about cold-start

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I populate a new GBrain knowledge base with my data?

Run the cold-start workflow after gbrain setup. It sequences imports by leverage: existing markdown first, then Google Contacts, Calendar, Gmail, conversation exports, and archives, with each phase gated on your consent.

What data sources can I import into GBrain first?

The priority order is markdown/Obsidian vaults, Google Contacts, Calendar history, Gmail threads, ChatGPT or Claude conversation exports, X/Twitter archives, file archives, and meeting transcripts. Each phase is independently valuable, so you can stop anytime.

Can I import Gmail and Calendar without giving the agent API tokens?

Yes. The default path on harnesses without ClawVisor is a Google Takeout export covering contacts CSV, calendar ICS, and Gmail mbox files. Direct OAuth is explicitly disallowed as a security risk.

What is ClawVisor and why does the import workflow use it?

ClawVisor is a credential gateway that vaults API keys, enforces task-scoped authorization, logs every call, and requires human approval for destructive actions. It prevents the agent from ever holding raw Google OAuth tokens.

How do I resume an interrupted brain import session?

Progress is tracked in ~/.gbrain/cold-start-state.json with completed phases and the next phase recorded. On restart, the workflow reads this file, skips finished phases, and resumes without repeating credential setup.

Does the import filter out newsletters and automated emails?

Yes. The Gmail phase auto-skips noreply senders, marketing mail, and notification bots, while conversation imports use a 1-5 significance scale and only keep conversations rated 3 or higher.