Highlights
- Adds an experimental-research profile with claim-evidence, figure, table, data, and model provenance templates.
- Adds safe adoption guidance and a read-only audit for mature workspaces.
- Adds dependency-free repository validation and Windows/Linux GitHub Actions.
- Adds citation metadata, version history, and a sanitized LLM inference optimization case study.
Verification
- 13 Node behavior tests passed.
- GitHub Actions passed on Windows and Ubuntu.
- Skill, YAML/CFF, UTF-8, line endings, and relative links were validated.
Compatibility
This release is backward compatible with the v1.0 competition scaffold. The research profile and workspace audit are opt-in.