This article provides a comprehensive introduction and functional analysis of ATG5 Conversion Workshop (Skills Edition) v3.09, a specialized technical refinery designed to convert AI procedures (Skills) across platforms according to the open Agent Skills (SKILL.md) standard, merge multiple procedures without redundancy, and split complex procedures into independent modules while preserving their shared backbone. The Skills Documentation Page link is provided within the article, and the Google Drive folder information containing the SKILL.md file that anyone can directly download and use on their own system is located at the end of the text.

Author: Aydın Tiryaki

As the artificial intelligence ecosystem transitions from legacy custom assistant structures to the open Agent Skills (SKILL.md) specification under the Agentic AI Foundation, ensuring that advanced procedures can be ported across diverse AI platforms with zero loss of fidelity has become essential. ATG5 Conversion Workshop (Skills Edition) is engineered to offload the heavy technical adaptation workload from the main ATG Gem Factory (v14.14) by executing cross-platform conversion, dynamic grouping/merging, and shared-backbone splitting directly on shared “Full Skills Texts”.

Technical Identity and Version Specifications

• Skills Name (EN / TR): ATG5 Conversion Workshop / ATG5 Dönüşüm Atölyesi
• Technical Registry Name (EN / TR): atg5-conversion-workshop / atg5-donusum-atolyesi
• Version and Lifecycle Status: v3.09 │ GemStat: 0 (Draft) │ Factory Version: v14.14
• Gem Repository Number: gdn057
• Character Distribution (EN): Title: 24 │ Description: 979 │ Instructions: 31,399 │ Total: 32,402 characters
• 📄Skills Documentation Page: https://aydintiryaki.org/2026/06/19/gdn057/

Architectural Structure and Operation 51 – Operation 54 Specialized Lines To maintain strict separation from the internal Mode and Task commands of the Factory and the operation numbers of other ATG workshops (ATG1: 11–12, ATG2: 21–28, ATG3: 31–32, ATG4: 41–43, ATG7: 71–72, ATG8: 81–82), all workflows within ATG5 Conversion Workshop are standardized under Operation 51 through Operation 54 in alignment with workshop number 5:

• Operation 51 (📜 General Rules and User Additions Protocol): Collects custom rules, general operating principles, and structural revisions requested by the user in listening mode and integrates them into the relevant sections of the loaded Full Skills Text with zero loss.
• Operation 52 (🔄 AI Platforms Skills / SKILL.md Conversion and Adaptation Engine): Converts Gemini Skills procedures into fully compliant SKILL.md packages tailored to the current specifications of ChatGPT/Codex Skills, Claude Skills / Claude Code, Grok, and Meta AI. It structures mandatory and optional YAML frontmatter fields (name, description, license, compatibility, metadata.short-description, allowed-tools), manages progressive disclosure token budgets and platform-specific truncation thresholds (such as the 200-character web upload limit on claude.ai versus the 1,024-character open standard limit, and the 5,000-token auto-compaction retention rule in Claude Code), and compiles a three-evidence-level Platform Skills Technical Limit Report prior to output.
• Operation 53 (🧩 Dynamic Skills Grouping and Multi-Merging): Offers a strategic solution for staying within platform active-procedure limits (such as the 100-enabled-skills limit in Gemini Skills or context discovery budgets in Codex and Claude) by combining two or more procedures. Users can choose between Method A (full integration under a single unified menu) and Method B (preserving each procedure’s independent task and menu hierarchy while deduplicating shared verification rules, core protocols, and process tracking lines).
• Operation 54 (✂️ Shared Backbone Protected Skills Splitting and Separation): Splits large or multi-domain procedures into independent Full Skills Texts while automatically carrying over the mandatory shared backbone—including version structure, data fidelity locks, verification protocols, web navigation rules, living assistant guidance, and process tracking lines—to every resulting module.

Smart Fidelity Lock, Recall Mechanism, and Direct Open URL Guarantee Every Full Skills Text loaded into the workshop undergoes deterministic character counting across its Title, Description, and Instructions sections, followed by a two-loop cross-verification upon output to prevent any unauthorized pruning. Through its Recall and Update Protocol, the workshop proactively identifies outdated conventions in loaded procedures and aligns them with the latest Agent Skills standards, while ensuring that every generated URL opens its target destination directly as a literal flat string without search engine redirection.

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/1OvzBe4AynhssEH1pAhFzX66TQXB_tFU6

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 architectural design and conceptual framework of the ATG5 Conversion Workshop (Skills Edition) v3.09 procedure introduced in this article, as well as all content, approaches, and views expressed herein, belong entirely to Aydın Tiryaki; the editorial assistance and writing support for this article were carried out by the active Gemini Pro model and Standard thinking level in accordance with ATG Gem Factory (Skills Edition) v14.14, ATG7 Drive Workshop v01.07, and ATG8 Document Workshop v01.18 standards, completed on 02.11.2026 08:25.