crimeacs

Claude Ingest Skill — Development skill for Claude Code

Development community

Claude Code skill: Ingest documents into atomic knowledge notes with semantic deduplication.

How to install Claude Ingest Skill

This entry records only its repository, not the path inside it, so there is no exact command to give. Open crimeacs/claude-ingest-skill and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Claude Ingest Skill does

Claude Code skill: Ingest documents into atomic knowledge notes with semantic deduplication.

Alternatives in Development

  • Portaljs Add Geo — Auto-ingest a geospatial file (GeoJSON, Shapefile, GeoPackage, KML/KMZ, FlatGeobuf, CSV-with-geometry) into a 2.3k ★
  • Brain — Set up or use your Professional Brain — init, ingest, recall, or review 1.3k ★
  • Opencove — Your infinite canvas workspace for agents, tasks, knowledge, and research 1.2k ★

README

claude-ingest-skill

A Claude Code skill that transforms documents into **atomic, interconnected concept notes** with semantic deduplication.

What Makes This Different

Unlike simple summarizers that create one page per document, this skill:

Traditional Summarizer This Skill
1 summary file per document 5-15 separate concept files
Concepts as bullet points Each concept is standalone note
No deduplication Semantic search finds similar notes
Content isolated per source Concepts accumulate across sources

**Result:** A growing knowledge graph where each concept links to others and gets richer as you add sources.

Installation

Claude Code (recommended)

/plugin marketplace add crimeacs/claude-ingest-skill
/plugin install ingest@crimeacs-claude-ingest-skill

Manual Installation

git clone https://github.com/crimeacs/claude-ingest-skill.git ~/.claude/skills/ingest

Usage

/ingest ~/Downloads/attention-paper.pdf
/ingest https://example.com/api-docs --internal
/ingest spec.docx --title "API Specification v2"
/ingest paper.pdf --dry-run

Arguments

Argument Description
[file-or-url] Path to document (.pdf, .docx, .md, .txt) or URL
--internal Create internal notes instead of literature notes
--title "..." Override the document title
--dry-run Preview extraction without writing files

How It Works

1. Extract Atomic Concepts

The skill analyzes your document and extracts 3-15 standalone concepts:

Input: attention-is-all-you-need.pdf

Extracted concepts:
- self-attention (technique)
- multi-head-attention (technique)
- positional-encoding (technique)
- transformer-architecture (finding)
- scaled-dot-product-attention (technique)

2. Semantic Deduplication

Before creating each note, the skill searches your vault for similar existing concepts:

Checking for duplicates...
✓