Evidence Runs validate runnable datasets
CSV/XLSX Dataset Cards are checked before queueing so metadata-only, unsupported, or unreadable sources fail early with clear next steps.
What's New
Epok v0.1.1 stabilizes Evidence Runs and project review flow.
This patch release validates runnable datasets before queueing, keeps completed run links in project context, and reduces default density across the core evidence workspace.
Completed run links stay in project context
Model Version, Evidence Log, and Evidence Pack actions now route to project-scoped review surfaces for the selected AI System.
Project pages are easier to scan
Overview, Evidence Pack, Dataset Card, and Evidence Run views now emphasize status, blocker, next action, and destination cues.
Release deploy path is chained
The release workflow now supports tag creation, GitHub Release publication, and production deployment from the same workflow run.
Release posture
Built for evidence review, not legal sign-off.
Epok helps teams collect, connect, and review AI System evidence. v0.1.1 keeps the language and workflow focused on review-ready evidence, source traceability, and reviewer decisions.