insights

Extract and categorize durable insights from historical transcripts into _insights.yaml.

4|Updated Feb 15, 2026

Search Tags:#documentation #insights #data analysis #transcripts #knowledge extraction #backfilling

Ready to run

npx skills add https://github.com/tandregbg/core-claude-skills --skill insights-tandregbg

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: insights
Source: https://github.com/tandregbg/core-claude-skills/tree/main/skills/insights
Command: npx skills add https://github.com/tandregbg/core-claude-skills --skill insights-tandregbg

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you systematically extract valuable, durable insights from your existing documents and transcripts, turning raw information into actionable knowledge.

Core Features & Use Cases

Retroactive Extraction: Processes historical transcripts and CLAUDE.md files to backfill insights into a structured _insights.yaml file.
Insight Categorization: Identifies and categorizes insights into types like decisions, preferences, learnings, opportunities, and patterns.
Use Case: After a series of project meetings, use this Skill to go back through all the transcripts and automatically identify all key decisions made, lessons learned, and recurring patterns, creating a consolidated knowledge base.

Quick Start

Run insights reprocess all to extract insights from all transcript files.

Frequently Asked Questions about insights

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

FAQPage Schema

How do I extract knowledge from historical transcripts automatically?▼

To extract knowledge from historical transcripts, this Skill processes your files retroactively to backfill durable insights into a structured _insights.yaml. It identifies and categorizes decisions, preferences, learnings, opportunities, and patterns from your existing data.

What is the best way to consolidate CLAUDE.md files into a single knowledge base?▼

Consolidating CLAUDE.md files into a single knowledge base is done by scanning them for durable insights and writing them into a centralized _insights.yaml file. This ensures data integrity and deduplication by tracking the original source file.

Can I categorize project decisions and learnings from past meeting transcripts?▼

You can categorize project decisions and learnings from past meeting transcripts by running the retroactive extraction process. It systematically identifies and groups historical data into categories like decisions, patterns, and opportunities.

How do I backfill insights from multiple documents without creating duplicates?▼

To backfill insights from multiple documents without duplicates, the extraction process ensures data integrity by deduplicating entries based on their source file. This prevents the same historical transcript from being processed twice.

Does transcript knowledge extraction require any external dependencies?▼

Transcript knowledge extraction requires no external dependencies to function. It operates independently to parse your historical documents and generate the consolidated _insights.yaml file.