From rubric-maker-skill
Produce provisional OSCE grade sheets from a Rubric Maker rubric and text analogues of OSCE evidence. Use when an agent needs to dry-run scoring against a draft rubric using post-encounter-note text, transcript text for audio evidence, observation-log text for video evidence, or a transcript-plus-observations bundle, with evidence, rationale, confidence, subtotals, totals, and deterministic grade-sheet validation. Do not use for generating student artifacts, revising rubrics, adjudicating clinical validity, or inspecting raw audio/video.
How this skill is triggered — by the user, by Claude, or both
Slash command
/rubric-maker-skill:grading-dry-runThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- Provisionally grading an example student artifact against an existing rubric.
scripts/validate_grade_sheet.py.generate-student-artifact.osce-rubric-transform.evaluate-dry-run.content-validation.rubric-import.note: written post-encounter-note text for Mode: note rows.transcript: spoken encounter transcript for Mode: audio rows.observation_log: timestamped observed actions for Mode: video rows.transcript_plus_observations: transcript text plus observation-log rows for mixed
Mode: audio and Mode: video grading.references/rubric-schema.md and references/grade-sheet-contract.md.references/evidence-rules.md for artifact-specific evidence boundaries.references/scoring-rules.md for scoring and total rules.mode.references/grade-sheet-contract.md.python3 scripts/validate_grade_sheet.py <grade-sheet-file>.items row per rubric item.score, max_score, evidence, rationale, and
confidence.unscorable: true, max_score, and
unscorable_reason.mode, only score it from a compatible text analogue: note mode from
note text, audio mode from transcript text, and video mode from observation-log
evidence.artifact_type: transcript_plus_observations when one dry run includes both audio
transcript evidence and video observation evidence.subtotals, total_score, max_score, and percentage when the user
requests a full grade sheet.
These totals are provisional scored-evidence totals that omit unscorable rows.unscorable_count and unscorable_max_score when any rows are unscorable.grading_friction_notes; do not rewrite the
rubric in this skill.references/rubric-schema.md: rubric field schema and score-anchor conventions.references/grade-sheet-contract.md: required grade-sheet shape and examples.references/evidence-rules.md: note, transcript, and observation-log evidence rules.references/scoring-rules.md: score selection, unscorable handling, and total math.npx claudepluginhub jamiesonlabutsw/maples-toolkit --plugin rubric-maker-skillCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.