ctaxnagomi/corpuslib-topics

CORPUSLIB Topics Dataset

CORPUSLIB — Agentic Corpus Library for Indirect Learning

This dataset contains the topic catalog for DeckerGUI's CORPUSLIB system. CORPUSLIB is a link-gated knowledge library focused on indirect learning as the AI/agentic technology space evolves.

Purpose

• Fallback system: When main learning sources are unavailable or undergoing maintenance, CORPUSLIB provides backup topic links
• Agent training: Structured topic data for training agentic fetch agents (dgui_corpuslib_fetch)
• Knowledge graph: Topic relationships and categories for graph engineering
• Corpus memory: Store and retrieve agent knowledge
• Skill library: Reference for agent skill development

Schema

Field	Type	Description
id	int	Unique topic identifier
topic	string	Topic name
category	string	Category classification
url	string	Direct link to original content
source	string	Source platform (x.com, youtube, etc.)
tags	array	Topic tags for search
status	string	Topic status (active/inactive)
added_by	string	Original contributor

Categories

• mobile-agents — Mobile agent development
• agent-frameworks — Agent frameworks and harnesses
• engineering-patterns — Loop, graph, memory, eval engineering
• ai-models — AI model discussions
• tools-ecosystem — MCP servers, GitHub repos, skills
• knowledge-management — Second brain, knowledge graphs
• research — Academic research and papers
• career — AI engineer career guidance
• education — Terminology and basics
• security — TLS, encryption
• training — Fine-tuning, model training

DeckerGUI Integration

This dataset connects to the DeckerGUI Agentic Ecosystem:

Services

• KPI Tokenizer (port 3004): Enterprise tracking, usage logging, dcf_agent progression
• DGUI Emitter (port 3006): Event emission, real-time updates, CompactDOM capture
• CORPUSLIB Service (port 3012): Standalone corpus library API
• Tunnel: https://corpuslib.deckergui.my

Architecture

• CORPUSLIB HUB (Public): Any AI agent can fetch topics and fill timetable
• CORPUSLIB INSTRUCT (Private): DeckerGUI embedded agents only (requires DGUI Emitter + DGM Compliance)

Agent Rules

For Non-DeckerGUI Agents (HUB Access)

1. Must use CompactDOM-only capture (snapDOM format)
2. Must fill the timetable before accessing topics
3. Must credit owner at end of applied codebase
4. No ClockIN-ClockOUT tracking required

For DeckerGUI Embedded Agents (INSTRUCT Access)

1. Must have DGUI Emitter (Digital API Key)
2. Must have DGM Compliance (DLIM Assessment)
3. Must ClockIN-ClockOUT for progression tracking
4. Full DOM capture available (optional)

Related Skills

• CORPUSLIB Skill: .agents/skills/corpuslib/SKILL.md
• Multi-Agent PR Review: .agents/skills/multi-agent-pr-review/SKILL.md
• Factory Missions: .agents/skills/factory-missions/SKILL.md

Corpus Memory

• CORPUSLIB DGUI: research_corpus/CORPUSLIB DGUI/ — Topic catalog and agent instructions
• Multi-Agent Review Corpus: docs/MULTI_AGENT_REVIEW_CORPUS.md — Knowledge corpus for PR review

Usage

Load Dataset

from datasets import load_dataset

dataset = load_dataset("ctaxnagomi/corpuslib-topics")

Filter by Category

engineering_topics = dataset.filter(lambda x: x["category"] == "engineering-patterns")

Search by Tags

graph_topics = dataset.filter(lambda x: "graph" in x["tags"])

License

DeckerGUI Ecosystem — MIT License

Contributing

To contribute topics, email ctaxnagomi@gmail.com

Citation

@dataset{corpuslib_topics_2026,
  title={CORPUSLIB Topics Dataset},
  author={DeckerGUI},
  year={2026},
  url={https://huggingface.co/datasets/ctaxnagomi/corpuslib-topics}
}