ghchinoy/mizan-authoring
Author metric-template packs for Mizan from an agent: scaffold a pack, create rubric/pointwise/pairwise/custom_schema templates via the mizan CLI, validate the pack creds-free (the CI PR gate), and run the export->PR->import collaborator loop; and generate a rubric from a brand book (suggest-first, agent-side synthesis) then freeze it into the registry. Use when you need to author, package, validate, or share a Mizan metric template or template pack, or to generate evaluation rubrics from a brand book or guidance document.
Author a Mizan metric template (or a whole template pack), validate it creds-free with mizan pack validate (the CI PR gate — it reads the exit code and text report, not JSON), and run the export→PR→import collaborator loop against a shared templates repo, driving the mizan CLI. Use when a user asks to author, scaffold, package, validate, or share a Mizan metric template, rubric, or template pack, or to prepare a pack for a pull request.
Discover starter/industry metric templates from the community mizan-templates repo (or any pack source / git URL) and import them into the local Mizan registry, using the mizan CLI over -o json. Preview with --dry-run, reconcile id clashes with --strategy, and browse what landed with registry list/get. Use when onboarding, when a user wants ready-made evaluation templates instead of authoring from scratch, or asks to pull in / update templates from a repo.
Turn a brand book or guidance document into a reviewed, frozen Mizan rubric metric template — the agent reads the doc, suggests existing templates to reuse, drafts sample prompts and candidate criteria, generates a draft rubric with mizan rubric generate, unions in hand-authored criteria, then validates and freezes it into the registry via mizan registry import, driving the mizan CLI. Use when a user asks to build, generate, or derive Mizan evaluation rubrics/criteria/metrics from a brand book, style guide, brand guidelines, content standards, or any guidance document, or to auto-suggest and author evals from written brand/quality guidance.