geometry-notebook-qc
Quality checks for educational geometry notebooks.
Install
mkdir -p .claude/skills/geometry-notebook-qc && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/14815" && unzip -o skill.zip -d .claude/skills/geometry-notebook-qc && rm skill.zipInstalls to .claude/skills/geometry-notebook-qc
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Quality-check visualization-first geometry notebooks. Use when Codex needs to review geometry notebooks, artifacts, and helper changes for standalone teaching quality, visual relevance, executable correctness, artifact integrity, stale paths, and dependency fit.Key capabilities
- →Check source grounding of notebook content
- →Verify standalone delivery of teaching material
- →Assess visualization quality in geometry notebooks
- →Evaluate library fit for appropriate tool usage
- →Check artifact integrity and path correctness
How it works
The skill reviews geometry notebooks for standalone teaching quality, visual relevance, executable correctness, and artifact integrity by checking source grounding, visualization quality, and library fit.
Inputs & outputs
When to use geometry-notebook-qc
- →QA a geometry notebook chapter
- →Check artifact consistency
- →Validate executable notebook cells
About this skill
Geometry Notebook QC
Purpose
Use this skill after authoring or revising a geometry notebook chapter. The goal is to catch weak standalone explanations, generic generated sameness, decorative visuals, missing artifacts, execution failures, stale paths, source-map gaps, and dependency/library mistakes.
Read references/geometry-library-catalog.md when evaluating whether a notebook uses the available stack appropriately.
Chapter-First Rule
- Inspect the course
AGENTS.md, assigned source map/span, changed notebook, helper changes, generated artifacts, and relevant scripts before judging the work. - A canonical teaching notebook must read as a direct chapter-specific lesson, not as a generic script-generated template.
- Scripts may support indexing, auditing, validation, artifact helpers, and small reproducible assets.
- Scripts must not mass-populate teaching notebooks with repeated generic markdown/code cells.
- Bootstrap scripts may create folder skeletons only unless the user explicitly asked for a rough bootstrap draft.
- Copyright safeguards remain required: no copied textbook prose, long exercise text, screenshots, page crops, figures, or page layouts.
Review Workflow
- Check source grounding:
- source span or source map is present
- chapter-specific terms, definitions, constructions, examples, and proof moves appear
- the reader can learn the chapter without opening the textbook
- Check standalone delivery:
- motivation, translation guide, examples, pitfalls, checks, labs, and takeaways are present where appropriate
- useful existing work was preserved during improvement passes
- prose near each visual explains what to inspect
- Check visualization quality:
- visuals are concept-specific, not decorative or quota-filling
- proof-heavy sections use dependency views, diagrams, symbolic experiments, finite models, deformation/limit views, counterexamples, or invariant trackers when useful
- labels, scale, aspect ratio, color, legends, and annotations make the geometry inspectable
- major concepts with no visual have a documented reason
- Check library fit:
- notebook includes chapter-specific library routing or clear library rationale
- Matplotlib-only treatment is flagged when the chapter calls for 3D mesh/surface, topology/TDA, projective/CV, geometric algebra, manifold/statistical, transport, or GIS tools
- optional/external tools are documented rather than silently imported
- Check artifacts:
- outputs live under the matching book-local
artifacts/subtree - filenames name the concept, not the renderer
- artifacts are displayed inline or linked near the relevant prose
- paths are current, relative/book-local, and not stale root paths
- artifacts have nonzero size and appropriate formats: PNG/SVG/HTML plus JSON/CSV checks when useful
- outputs live under the matching book-local
- Check computation:
- final cells assert core symbolic, numeric, or geometric invariants
- JSON/CSV summaries record reproducible checks when useful
- notebook executes with the narrowest relevant course validation command
Anti-Generic Audit Guidance
When several chapter notebooks are in scope, check for:
- repeated markdown shingles across unrelated chapters
- repeated code-cell fingerprints or identical setup-to-display blocks
- identical notebook outlines with only chapter names changed
- repeated artifact names such as
concept-dependency-map.pngwithout chapter-specific companions - missing chapter-specific terms from headings, markdown, checks, and artifact names
- missing library-routing notes
- visuals with no nearby explanatory markdown
- notebooks driven primarily by one monolithic
generate_*orauthor_*script - shared helper calls that hide the chapter pedagogy, such as one
create_chapter_visuals(...)cell standing in for the lesson
Treat these as warnings first, then fail the work when they make the notebook generic, decorative, or not standalone.
Command Guidance
Use the commands listed in each course AGENTS.md. For quick checks, prefer:
uv run python -m compileall -q <course>/utils <course>/scripts- course notebook audit script
- course visual audit script when present
- limited
validate_*_course.pyexecution before full validation git diff --check
Run full notebook validation when shared utilities, setup cells, artifact helpers, or many notebooks changed.
Final QC Report
Return a compact report with:
pass/fail summarystandalone teaching issuesvisualization issuesartifact issuesexecution issuesgeneric-generation warningsrecommended fixes
Lead with actionable findings. For each issue, include the file path, tight location when possible, why it weakens the notebook, and the concrete fix. If there are no findings, state that and list checks not run.
When not to use it
- →The user needs to mass-populate teaching notebooks with generic content
- →The user wants to copy textbook prose or figures
- →The user needs a general-purpose code linter
Limitations
- →Does not mass-populate teaching notebooks with repeated generic markdown/code cells
- →Does not allow copying textbook prose, long exercise text, or screenshots
- →Treats repeated markdown shingles or identical notebook outlines as warnings first
How it compares
This skill provides a specialized quality check for geometry notebooks, focusing on pedagogical effectiveness and visual accuracy, unlike a generic code review or content audit.
Compared to similar skills
geometry-notebook-qc side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| geometry-notebook-qc (this skill) | 0 | 3mo | No flags | Intermediate |
| test-knowledge | 0 | 4mo | No flags | Advanced |
| tutorial-engineer | 5 | 4mo | No flags | Intermediate |
| learning-path-structure-review | 0 | 1mo | No flags | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by Rah-Rah-Mitra
View all by Rah-Rah-Mitra →You might also like
test-knowledge
uchicago-bfi-gnlab
Generates a quiz to test the user's knowledge of a codebase or script. Supports two modes: script-specific (deep dive on one script, invoke with a script path) and whole-project (architecture, design choices, pipeline, invoke with `project`). Reads actual source files, git history, and directory str
tutorial-engineer
sickn33
Creates step-by-step tutorials and educational content from code. Transforms complex concepts into progressive learning experiences with hands-on examples. Use PROACTIVELY for onboarding guides, feature tutorials, or concept explanations.
learning-path-structure-review
ArmDeveloperEcosystem
Review Arm Learning Paths for structural soundness, task ownership, scope discipline, progression, validation, further reading placement, recap sections, and performance-learning integrity. Use when asked to review, create, reorganize, or assess the structure of content under content/learning-paths.
exam-construction
psychelzh
>
write-concept
leonardomso
Write or review JavaScript concept documentation pages for the 33 JavaScript Concepts project, following strict structure and quality guidelines
opik-university
j-mckerracher
>