skills.rest
AMC-JTC
AMC-JTCCOMMUNITY
Explore all skills by this author →
GitHub

ingest

Routes meetings, articles, media, and documents into a structured knowledge graph with citations and back-links.

Updated Aug 17, 2026
Search Tags:#knowledge graph#citations#ingestion#entity extraction#meeting transcripts#personal knowledge management#back-links
One-click install
npx skills add https://github.com/AMC-JTC/gbrain-1 --skill ingest-amc-jtc
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ingest
Source: https://github.com/AMC-JTC/gbrain-1/tree/main/plugin/skills/ingest
Command: npx skills add https://github.com/AMC-JTC/gbrain-1 --skill ingest-amc-jtc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Raw content like meeting transcripts, articles, podcasts, and social posts piles up without structure, making it impossible to recall who said what, when, and why it matters. This Skill ingests that content into a personal knowledge brain, extracting entities, preserving sources, and wiring everything into a queryable knowledge graph. ## Core Features & Use Cases - Multi-format ingestion: Processes meeting transcripts, articles, videos, podcasts, PDFs, screenshots, and social media posts with format-specific workflows. - Entity detection and back-linking: Detects people, companies, and concepts on every message, creates or updates their pages, and enforces bidirectional links between all mentioned entities. - Citation and provenance enforcement: Requires inline [Source: ...] citations on every fact and preserves raw sources via size-routed upload (git for small files, cloud storage for large media). - Use Case: After a meeting with a startup founder, paste the transcript and the Skill creates a meeting page, updates the founder's and company's pages, adds timeline entries to all three, and stores the raw transcript for provenance. ## Quick Start Ingest this meeting transcript into the brain and update all mentioned people and companies.

Frequently Asked Questions about ingest

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

FAQPage Schema
How do I ingest a meeting transcript into a knowledge base?▼

Provide the full transcript and the Skill creates a meeting page with analysis, updates every attendee's and company's State section, appends timeline entries to each entity, and preserves the raw transcript. A meeting is not fully ingested until all entity pages are updated.

How do I save articles and web content with proper citations?▼

Share the URL and the Skill fetches the content, extracts title, author, and key arguments, then writes a page filed by primary subject with an inline [Source: publication, URL, date] citation. The raw source is preserved for provenance.

Can I ingest YouTube videos and podcasts into a knowledge graph?▼

Yes, provide a URL or local media file and the Skill obtains a speaker-diarized transcript, saves the raw transcript, extracts key ideas and verbatim quotes with real speaker names, and back-links all mentioned entities. Every media page must link to its raw transcript.

What happens to large files when preserving raw sources?▼

Files under 100 MB stay in git in .raw sidecar directories, while files 100 MB or larger and all media are uploaded to cloud storage via TUS resumable upload, leaving a .redirect.yaml pointer in the brain repo.

Why does the Skill rewrite State sections instead of appending?▼

State sections hold the current best understanding of an entity, so appending would create stale and contradictory content over time. New information is merged by rewriting the State section, while timeline entries capture the historical record.

When should I test before bulk ingestion?▼

Always test on 3-5 items before batch processing videos or meetings. Read the actual output for quality issues like generic titles or missing back-links, fix the approach, then bulk execute with commits every 5-10 items.

Related Skills

Explore More Personal & Entrepreneur Skills→
kulharir7kulharir7

summarize

Summarize text and transcripts from URLs, podcasts, and local files.

Community
Intermediate
PostergullyPostergully

query

Query a knowledge base with 3-layer search and citation-backed answers.

Community
Advanced
yunusgungoryunusgungor

meeting-ingestion

Parse meeting transcripts into brain pages with attendee enrichment and timeline merge.

Community
Advanced
cardox6cardox6

neo4j-document-import-skill

Ingest unstructured documents into Neo4j as knowledge graphs using LLM extraction.

Community
Advanced