BUILD

Source Code and Whitepaper

Read the design. Explore the source. Build with the format.

Start Here

  • Source Code

    The Python reference implementation, specification, JSON Schemas, examples and tests in the public GitHub repository.

    Python 3.10+ · Standard library only · Apache 2.0

  • Whitepaper

    The collection format, provenance, AI exchange and security boundaries, with diagrams and references.

    PDF · LaTeX · Mermaid diagrams

This website follows a reviewed repository revision for specification 1.0.0. GitHub contains the source and its development history.

Run the Source

git clone https://github.com/liminalatoms/showpapers-format.git
python3 showpapers-format/tool/showpapers_format.py validate your.showpapers

Creating and editing need Linux or macOS. Commands that create files require a new output path and refuse to overwrite an existing file.

Developer Quick Start · Schema Reference · Examples

For AI Agents

The Agent Guide explains how to create and update a collection. Give an assistant llms.txt as its starting point, or use the complete text edition.

For an assistant that supports local skills, download the workspace skill. It bundles the instructions and reference tool in the expected folder layout; it does not add another protocol implementation.

Inspect the ZIP before installing it. For Codex, extract it into .agents/skills from your workspace root:

mkdir -p .agents/skills
unzip -n /path/to/showpapers-collections-skill.zip -d .agents/skills

The extraction command keeps existing files. For a fresh installation, use an empty skill folder. Ask your assistant to use the showpapers-collections skill and attach the papers or collection. See workspace examples. For another assistant, use its supported local skill directory.

The standalone writer remains available for constrained environments. Most developers should use the GitHub repository above.