What problem does it solve?
It reduces the effort readers face when trying to understand and reproduce scientific papers by packaging the paper with the instructions and artifacts needed to run an AI agent that can explain and help reproduce the work.
Core Features & Use Cases
- Staging an APP publication bundle: Creates a clean, self-contained publication candidate in publication-staging/ that an external agent can treat as the repository root.
- Multi-phase validation and paper-agent documentation: Runs structured validation, produces AGENTS.md/README.md/CLAUDE.md for the paper agent, and performs a staging-root smoke test.
- Controlled publishing workflow: In real mode, promotes the validated staging tree to a tagged public release and generates APP_PUBLICATION.json; in developer sandbox mode, records an implementation-test outcome without creating public compliance records.
- Use case: Turn an existing author working repo (paper + code/data/figures) into an APP release so readers can open the repo with an AI coding agent and ask questions grounded in the paper and reproducible artifacts.
Quick Start
Ask your AI assistant to run /publish-paper with your working repository as context, and follow the phase prompts until the validated publication-staging/ is ready for real publication or a developer sandbox test.