ShaderCells

ShaderCells provides resources for writing, debugging, and optimizing ReShadeFX (HLSL) shaders. This repository includes a specialized AI agent skill and a comprehensive Retrieval-Augmented Generation (RAG) knowledge base.

Agent Skills (./skills)

Extend AI CLI tools with the reshadefx-coder skill. Agent skills provide specialized capabilities that allow AI models to execute code, search files, and perform complex tasks autonomously. Learn more at the Anthropic Agent Skills site.

Installation

Copy the skills/reshadefx-coder folder to your CLI’s search path:

Claude Code

• Global Path: ~/.claude/skills/
• Local Path: ./.claude/skills/

Gemini CLI

• Global Path: ~/.gemini/skills/ or ~/.agents/skills/
• Local Path: ./.gemini/skills/ or ./.agents/skills/

Mistral Vibe

• Global Path: ~/.vibe/skills/
• Local Path: ./.vibe/skills/

Example Path:

./.claude/skills/reshadefx-coder

Chatbot Experts (./rag)

Turn popular chatbot applications into ReShadeFX experts using the provided RAG files. Retrieval-Augmented Generation (RAG) connects AI models to specific data sources, ensuring responses are accurate and grounded in the provided technical documentation.

1. Download Resources

1. Download the repository as a .zip from the Project Page.
2. Unzip the files to a known location.

2. Configure Your Assistant

Google Gemini (Gems)

1. Open Gemini and log in.
2. Select Gems > New Gem.
3. Paste rag/Description.txt into Description.
4. Paste rag/Instructions.txt into Instructions.
5. Upload all files from rag/knowledge/ to the Knowledge box.

Mistral AI (Le Chat)

1. Open Le Chat and log in.
2. Open the Intelligence tab and select Libraries.
3. Click New Library, name it “Shader Knowledge,” and upload all files from rag/knowledge/.
4. Open the Agents tab and select Create an agent.
5. Paste rag/Description.txt into Purpose and rag/Instructions.txt into Instructions.
6. Open the Knowledge tab and checkmark the “Shader Knowledge” library.

AnythingLLM (Local)

1. Open AnythingLLM and create a workspace named “ReShadeFX”.
2. Click the Upload icon, add all files from rag/knowledge/, and select Move to Workspace.
3. Open Chat Settings > System Prompt and paste rag/Instructions.txt.

GPT4All (Local)

1. Open GPT4All and navigate to Settings > LocalDocs.
2. Add rag/knowledge/ to the indexed folders.
3. In Settings > Model, clone your preferred model.
4. Paste rag/Instructions.txt into the System Message field.