A comprehensive introduction to the dual-mode AI Lexicon (Skills Edition) v1.18 design, which examines individual artificial intelligence models across all dimensions and cross-compares any AI feature across all known platforms; the product’s Skills Documentation Page link and Google Drive folder information containing the SKILL.md file that anyone can directly download and use in their own system are provided at the end of this article.

Aydın Tiryaki

The artificial intelligence ecosystem is no longer limited to simple text-generating chatbots; it has evolved into a vast operational layer featuring custom assistant architectures (Gems, Skills, Custom GPTs, Projects, Spaces), multi-million-token context windows, persistent memory systems, interactive code and document workspaces (Canvas, Artifacts), autonomous deep research agents (Deep Research), scheduled tasks, and Model Context Protocol (MCP) connectors. In this multi-layered landscape, users, researchers, and enterprises need both an end-to-end breakdown of any single AI model and a cross-platform comparison showing how a specific feature (such as Skills architecture or custom assistant naming) is implemented across all other known AI platforms. AI Lexicon (Skills Edition) v1.18 is a dual-mode technical, architectural, and commercial analysis guide designed in the persona of a Senior AI Analyst to unify both capabilities under a single roof.

Developed in accordance with ATG Gem Factory (Skills Edition) v14.16 standards in Türkiye, AI Lexicon v1.18 carries the gdn046 catalogue identifier and features a comprehensive architecture of 17168 characters in total (Title: 10, Description: 975, Instructions: 16183 characters). Whenever active in any conversation, the system utilizes its dynamic latest version trigger protocol to operate seamlessly with the latest installed version and presents every link as a verified, direct open URL without search engine redirection.

📄Skills Documentation Page: https://aydintiryaki.org/2026/06/19/gdn046/

Dual-Mode Operational Architecture: Mode 1 (Single AI/Domain Analysis) and Mode 2 (Cross-AI Feature Comparison) AI Lexicon v1.18 operates across two primary axes:

1. Mode 1 — Single AI / Model or Specialized Domain Analysis: When a user enters an AI name (Gemini, ChatGPT, Claude, DeepSeek, Grok, Copilot, Perplexity, Midjourney, Sora, etc.) or a specialized domain (Video, Text, Audio, Image, Code), the system reports the tool’s official identity, country of origin, direct web/desktop/mobile store links, history, industry benchmark scores (LMSYS Chatbot Arena, MMLU, HumanEval, SWE-bench, MATH), and 3-month/1-year version evolution.
2. Mode 2 — Feature, Capability, and Architecture Cross-Comparison: When a user queries a specific AI feature or structural concept rather than a single model (e.g., “How does the Skills architecture work across AI platforms?”, “What are custom assistants called in each AI?”, “Which platforms offer Deep Research or Canvas?”), the system scans all known AI platforms. For instance, it maps how custom assistant and modular skill structures are designated as Gems and Skills (SKILL.md) in Gemini, Custom GPTs and Projects in ChatGPT, Projects, Styles, and Skills in Claude, Spaces in Perplexity, Agents in Microsoft Copilot, and Bots in Poe, presenting their instruction limits, knowledge base capacities, and sharing permissions in both an emoji-numbered analytical list and a comparative matrix table.

Ten Strategic Architectural and Operational Analysis Layers With version 1.18, every single-model breakdown and cross-platform comparison incorporates 10 deep analytical dimensions:

1. Custom Assistant, Personalization, and Skills Architecture: Gem, Skills (SKILL.md), Custom GPT, Project, Space, and Agent structures, system instruction character ceilings, and store/sharing permissions.
2. Context Window, Output Ceiling, and Persistent Memory: Maximum input and output token/page capacities, text pruning tendencies on long outputs, Long-Term Memory across sessions, and Personal Context support.
3. Model Family, Sub-Models, and Reasoning Tiers: Fast/Economical (Flash, Mini, Haiku), Standard/Flagship (Pro, Sonnet, 4o), and Deep Reasoning (Extended Thinking, Deep Think, o1/o3, R1) sub-models and thinking level controls.
4. Interactive Workspace and Code Interpreter: Side-panel interfaces (Gemini Canvas, Claude Artifacts, ChatGPT Canvas) and background Python/VM Sandbox execution environments.
5. Multimodal Input/Output Matrix: Accepted input file formats and directly generated downloadable files (.docx, .xlsx, .pptx, .pdf, .md, .jpg, .mp4), plus real-time bidirectional voice/video modes (Live / Advanced Voice Mode).
6. Autonomous Agent, Deep Research, and Scheduled Tasks: Multi-step Deep Research report generation, subagent delegation, autonomous browser/computer actuation, and time- or event-triggered Scheduled Actions.
7. Ecosystem Connectors and MCP Support: First- and third-party integrations (Google Workspace, Microsoft 365, GitHub, Slack, Notion) and open-standard Model Context Protocol (MCP) support.
8. Core Architecture, Knowledge Cutoff, and Citation Discipline: Open-Weight versus Proprietary structure, training knowledge cutoff dates, live search engine infrastructure, and direct open URL grounding accuracy.
9. Safety Guardrails, Copyright, and Watermarking: Content filter sensitivity, commercial copyright terms, and invisible digital watermarking (SynthID, C2PA metadata).
10. Prompt Engineering Characteristics and Behavioral Scorecard: Adherence to long system instructions, chronic model habits (pruning, over-summarization, code truncation), and practical mitigation strategies.

Commercial Tiers, Verification, Quiz, and 9-Option Visual/Video Menu The system thoroughly documents Free, Pro/Plus, and Enterprise pricing, message/token/credit quotas, per-million-token API costs, GDPR/CCPA/EU AI Act privacy compliance, and regional availability. Every report concludes with direct open URLs to Wikipedia, authoritative technical sources, and YouTube videos/playlists. Dual-layer data verification (Hesitation Notice), a 5-question expertise quiz transferable to Google Forms, Summary/Detailed/Index Living Assistant guidance, and a 9-option visual/video generation menu complete the end-to-end research experience.

Google Drive Folder Link and SKILL.md File You can go to the folder below to download and directly use the SKILL.md file from this folder. If you cannot access this folder or file, it means the link is temporarily suspended because the development of the relevant Skills is currently in progress; please try again later.

📁 Google Drive Folder Path (EN): https://drive.google.com/drive/folders/1iXysPTfudLvNggovI7Dh9fHv7iukS111

Note: You can make your own developments and edits on the SKILL.md file you obtain from here. However, out of respect for the effort, we kindly ask you to keep the Skills creator and editor information included here.

The AI Lexicon (Skills Edition) v1.18 design introduced in this article was designed and developed by Aydın Tiryaki, and all content, analytical approaches, and views in the article belong to the author; editorial and writing assistance for this article was performed by the Gemini Pro model at the Extended thinking level within the ATG Gem Factory (Skills Edition) v14.16 and ATG6 Document Workshop (Skills Edition) v01.21 architecture (07.11.2026 22:23).

---

| aydintiryaki.org | YouTube | Aydın Tiryaki’nin Yazıları ve Videoları │Articles and Videos by Aydın Tiryaki | Bilgi Merkezi│Knowledge Hub  | ░ Virgülüne Dokunmadan │ Verbatim ░ | ░ Yapay Zeka Dağarcığı │AI Lexicon (Gemini Skills) ░ 07.11.2026

---