Working With AI

Install the AI skill

Download the Export Skafold Context skill and configure it in ChatGPT, Claude, or Gemini.

The Export Skafold Context skill converts a project conversation or brief into the versioned JSON package Skafold can import. The package includes the core instructions plus schema, node type, section preset, and example references.

Export Skafold Context skill

Includes the instructions, exchange schema, example package, node types, and section preset reference.

Version 1.2 · ZIP archive · 6 files

Extract the ZIP before configuring ChatGPT or Gemini. Claude can use the ZIP as a native custom skill.

Install in Claude

Claude supports packaged custom skills directly.

  1. Download the skill ZIP above.
  2. Open Claude and go to Settings.
  3. Open Capabilities or Skills. The label can vary by plan and interface version.
  4. Enable Skills if required.
  5. Choose the option to upload or add a custom skill.
  6. Upload export-skafold-context-v1.2.zip.
  7. Enable the skill.

Start a conversation with:

Use the Export Skafold Context skill to package this project for import into Skafold.

For an existing Skafold project, attach the JSON exported from Skafold and say:

Use this existing Skafold package as the baseline. Apply the changes from our conversation, preserve IDs for unchanged items, and return one complete revised JSON package.

Install in ChatGPT

ChatGPT GPTs do not install the folder as a native skill. Configure a custom GPT with the instructions and reference files.

  1. Download and extract the ZIP.
  2. In ChatGPT, open Explore GPTs and choose Create.
  3. Open the GPT configuration view.
  4. Name it Export Skafold Context.
  5. Open SKILL.md and paste its instruction content into the GPT Instructions field. You can omit the opening YAML block between the --- markers.
  6. Upload these files as Knowledge:
    • references/exchange-schema.json
    • references/example-package.json
    • references/node-types.md
    • references/section-presets.md
  7. Enable file generation or data analysis if the GPT configuration offers it.
  8. Save the GPT as private or share it with your team.

Use the GPT in a new chat, or type @ in an existing ChatGPT conversation and select the GPT when that option is available. Then ask it to export the current project context.

Install in Gemini

Gemini Gems use instructions and uploaded knowledge files rather than a packaged skill folder.

  1. Download and extract the ZIP.
  2. In Gemini, open the Gems manager and create a new Gem.
  3. Name it Export Skafold Context.
  4. Paste the instruction content from SKILL.md into the Gem instructions. Omit the opening YAML block.
  5. Add the four files from the references folder as Gem knowledge.
  6. Save the Gem and begin a chat with it.

Ask:

Package this project conversation as one validated Skafold context JSON file.

What the files do

  • SKILL.md defines the workflow and output rules.
  • exchange-schema.json defines the accepted JSON structure.
  • example-package.json demonstrates complete field usage.
  • node-types.md explains where different kinds of project context belong.
  • section-presets.md lists recognised Skafold page section names.
  • agents/openai.yaml provides optional agent metadata; ChatGPT and Gemini do not require it.

Keep the skill current

The version is recorded in generator.skillVersion. Download a fresh copy after Skafold announces a schema or skill update, then replace the prior skill or knowledge files in your AI tool.