cos-extract-from-sources

Extract facts from email, documents, transcripts, calendar, and web pages daily.

Updated Jun 5, 2026

Search Tags:#data privacy #fact extraction #memory isolation #daily cadence #restricted profile

Ready to run

npx skills add https://github.com/fontesgerards/chief-of-staff --skill cos-extract-from-sources

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: cos-extract-from-sources
Source: https://github.com/fontesgerards/chief-of-staff/tree/main/engine/skills/cos-extract-from-sources
Command: npx skills add https://github.com/fontesgerards/chief-of-staff --skill cos-extract-from-sources

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill automates the extraction of facts from email, documents, transcripts, calendar, and web pages, without writing memory, ensuring data privacy and security.

Core Features & Use Cases

Read-Only Extraction: Extracts facts from various sources without modifying the AI's memory.
Daily Cadence: Runs daily at a specific time, providing a consistent flow of extracted information.
Isolation: Operates in a restricted profile, preventing unauthorized writes to memory.
Consumption by Other Skills: Supplies extracted data to other skills for further processing.

Quick Start

Use the cos-extract-from-sources skill to automatically extract facts from your email and documents.

Frequently Asked Questions about cos-extract-from-sources

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

FAQPage Schema

How do I extract facts from emails and documents without writing to AI memory?▼

To extract facts from documents and emails without writing to AI memory, use a restricted profile that operates in strict isolation. This read-only approach ensures data privacy by pulling information from sources while preventing any unauthorized memory writes.

How do I automate daily fact extraction from transcripts and web pages?▼

You can automate daily fact extraction from transcripts and web pages by setting up a scheduled daily cadence. This process automatically pulls data from various sources at a specific time each day, providing a consistent flow of extracted information for downstream use.

Can I use extracted facts directly for processing or do they need another skill?▼

Extracted facts cannot be used directly for complex processing and must be consumed by other skills. The extraction process focuses solely on pulling data in a read-only format, supplying the results to downstream skills for further analysis and handling.

Does a restricted profile ensure data privacy when extracting facts from a calendar?▼

Yes, a restricted profile ensures data privacy when extracting facts from a calendar. By operating in an isolated environment, the extraction process prevents unauthorized writes to memory, securing sensitive schedule information while gathering required facts.

What are the limitations of using a read-only profile for fact extraction?▼

The main limitation of using a read-only profile for fact extraction is the inability to write or persist information directly to AI memory. This means the extracted data must be consumed by other skills immediately or stored externally for further processing.