Skip to content

Rubric Maker Skills

Use rubric-maker-skill when you need to create, import, review, transform, validate, or test OSCE rubrics and related grading artifacts.

It is the first installable plugin in the UT REAL Project MAPLES Plugin Marketplace. The package is available for Codex CLI and Claude Code.

  • post-encounter-note-rubric: Draft a Mode: note post-encounter-note grading rubric from case files, station instructions, SP scripts, sample notes, checklists, or clinical scenarios.
  • rubric-import: Import existing rubric source material into Rubric Maker YAML or JSON while preserving source wording and structure.
  • osce-rubric-transform: Adapt or enhance OSCE rubrics for new cases, contexts, scoring scales, missing fields, or style templates.
  • osce-rubric-review: Review OSCE rubrics for clinical alignment, safety, observability, objectivity, feasibility, reliability, and scoring clarity.
  • content-validation: Validate rubric issues through multi-perspective critique and final recommendations.
  • generate-student-artifact: Generate synthetic student notes, encounter transcripts, or transcript-derived observation logs from a case and learner performance profile.
  • grading-dry-run: Trial-grade a note, transcript, observation log, or evidence bundle against an OSCE rubric and produce an evidence-backed grade sheet.
  • evaluate-dry-run: Analyze a case, rubric, student artifact, and trial grade sheet to identify prioritized rubric improvements.
FieldValue
Package/plugin namerubric-maker-skill
Version0.1.0
PublisherUT REAL Project MAPLES
LicenseLicenseRef-UTSW-Academic-Research-Only
Supported runtimesCodex CLI and Claude Code

The license is summarized here as academic research use only and non-commercial. Review the repository license terms before distribution or use outside an academic research context.

The plugin provides standalone skill guidance and bundled script support for importing, drafting, reviewing, transforming, validating, rendering, generating synthetic artifacts, and dry-run grading/evaluation.

It does not itself provide raw audio or video inspection, persistence for assessments and grading jobs, live simulator sessions, or a full multi-model validation runtime with persistence and streaming. Those capabilities require an external runtime or application layer.

Tutorial, how-to, and reference sections are under development.