Marble Skill Taxonomy

An open, structured taxonomy of what children learn across the primary/elementary years — decomposed into fine-grained "micro-topics", wired into a prerequisite graph, and aligned to national curriculum standards. Produced by Marble.

Version: v1 · Topics: 1,930 · Prerequisite edges: 4,102 · Subjects: 9

See it

Every dot is a micro-topic, colored by subject; height is age; each thread is a prerequisite (full-quality video). Explore it interactively at withmarble.com/curriculum — tap any concept to trace everything a learner must master before it.

What this is

Most curriculum data is either a flat list of standards or locked inside a product. This dataset is a connected graph of learning:

• 1,930 micro-topics — a single, teachable idea (e.g. "Building sentences", "Apparent brightness of stars"), each with a plain-language description, mastery evidence criteria, a type (conceptual / procedural / representational / language / meta), a subject + domain, and an approximate age range.
• 4,102 prerequisite dependencies — a directed acyclic graph: "topic X depends on prerequisite Y", each edge tagged hard/soft and carrying a one-line reason.
• Curriculum alignment — each micro-topic links to the standards it was distilled from (NGSS, Common Core, the UK National Curriculum, and more).
• Teaching research — 1,049 records covering Science and Mathematics topics: teaching approaches, concrete learning activities, feedback, assessment guidance and linked sources. Coverage is partial; the other 881 topics have no research in this release.
• Domain clusters — 176 parent-friendly one-paragraph summaries per (subject, domain, age band).

Subjects

Subject	Topics
Science	595
Mathematics	587
English	368
History	112
Geography	25
Personal & Social Development	124
Life Skills	42
Computing	59
Learning to Learn	18

Files

All data lives in data/ as UTF-8 JSON. See schema/ for JSON Schemas and manifest.json for counts + SHA-256 checksums.

File	What it holds
data/topics.json	The micro-topics (graph nodes).
data/dependencies.json	Prerequisite edges (topicId depends on prerequisiteId).
data/curriculum-standards.json	The source curriculum standards, grouped by curriculum.
data/clusters.json	Parent-friendly domain summaries.
data/teaching-research.json	Teaching research joined by topicId, with explicit coverage and publication notes (~24 MB).
data/archived-teaching-research.json	Historical guidance for merged or refocused topics, with original records and current topic associations.
data/manifest.json	Counts, per-subject breakdown, per-file checksums.

The standards catalogue includes 620 Australian Curriculum v9 Foundation–Year 6 content descriptions (acara-v9), bringing it to 8 sources / 3,881 standards. The catalogue spans English, Mathematics, Science, HASS, Health and Physical Education, Digital Technologies, and Design and Technologies. All 593 content descriptions in the agreed scope are fully covered, as recorded in the scope coverage summary. Health coverage targets the Personal, social and community health strand; the 27 PE standards are excluded. Arts, Languages, elaborations and achievement standards are outside this contribution. This is coverage of the listed learning areas, not the entire Australian Curriculum.

Australian Mathematics has 442 topic alignments across 325 topics. All 139 Foundation–Year 6 content descriptions are fully covered in the coverage report, following gap authoring and review of practical work, virtual materials, digital tools and Australian currency.

Australian English has 630 topic alignments across 267 topics. All 189 Foundation–Year 6 content descriptions are fully covered in the coverage report, following gap authoring and review of Australian literature, phonics, comprehension, digital texts, multimodal creation and spoken presentations. This batch adds 82 English topics.

Australian Science has 245 topic alignments across 131 topics. All 77 Foundation–Year 6 content descriptions are fully covered in the coverage report, following gap authoring and review of practical investigations, measurements, data, scientific communication and Australian contexts. This batch adds 24 Science topics.

Australian HASS (Humanities and Social Sciences) has 210 topic alignments across 124 topics. All 89 Foundation–Year 6 content descriptions are fully covered in the coverage report. HASS is a curriculum grouping mapped onto the shared History, Geography, Personal & Social Development and Life Skills subjects. The authoring review records the reuse, topic splits and subject/domain decisions.

Selected Australian Health content has 117 topic alignments across 79 topics. All 32 Foundation–Year 6 Personal, social and community health standards are fully covered in the coverage report. The authoring review records reuse, practical evidence and review corrections. Health uses the existing Personal & Social Development and Science subjects. The 27 PE standards remain excluded; these figures do not represent full Health and Physical Education coverage.

Australian Digital Technologies has 93 topic alignments across 74 topics. All 38 Foundation–Year 6 content descriptions are fully covered in the coverage report. The authoring review records reuse and domain decisions. Computing now includes shared domains for digital systems, data, programming, digital design, collaboration and privacy.

Australian Design and Technologies has 111 topic alignments across 78 topics. All 29 Foundation–Year 6 content descriptions are fully covered in the coverage report. The authoring review records shared-subject placement, reuse and the final factual corrections. Physical making uses Science; food preparation uses Life Skills; food and fibre production uses Geography.

See PROVENANCE.md for sources and attribution.

Adding curricula

Curricula map onto one shared taxonomy. Keep subjects and domains consistent across countries; a curriculum's learning-area labels do not create separate subject trees.

• Reuse a topic when it teaches the same concept or skill, and extend it where needed to cover another curriculum's requirements.
• Reuse existing domains before adding a distinct new domain. Keep names consistent and avoid parallel country-specific domain trees.
• Keep each topic focused on a teachable concept or skill. A standard can link to several topics; its bundled requirements need not become one topic.
• Add country-specific topics where the knowledge genuinely differs, such as a country's history or constitutional arrangements. Use shared topics for common concepts and skills, with local examples where useful.

A topic

{
  "id": "mt_N8CpN1EJrP",
  "type": "CONCEPTUAL",
  "subject": "English",
  "domain": "Grammar & Punctuation",
  "name": "Building sentences",
  "description": "Understand that words combine to make sentences — a sentence expresses a complete thought…",
  "ageRangeStart": 4,
  "ageRangeEnd": 6,
  "centrality": 0.257,
  "evidence": [
    "Distinguish between complete sentences and fragments",
    "Compose a complete sentence with a subject and verb"
  ],
  "assessmentPrompt": "If {{name}} says something like \"The dog\", can they tell you that's not a complete sentence…?",
  "standards": ["ccss-ela:L.K.1f", "uk-nc-2013:Eng.App2.Y1.Sent.1"]
}

• id — stable identifier (mt_…), referenced by dependencies and by neighbours.
• standards — keys into curriculum-standards.json ("<curriculum-slug>:<code>").
• assessmentPrompt — a natural-language check for the idea. Contains a {{name}} placeholder (the child's name); substitute or strip before display.

A dependency

{ "topicId": "mt__00ZSLnB7p", "prerequisiteId": "mt_VBl1T1sFCM", "strength": "hard",
  "reason": "Must understand vibrations make sound before finding volume patterns" }

topicId depends on prerequisiteId. Reverse the edge to get "unlocks".

Teaching research

data/teaching-research.json adds how a topic might be taught to the graph of what children learn. It covers 547 Science and 502 Mathematics topics. coverage.missingTopicIds and per-subject counts make the 881 topics without research explicit; absence is not a judgement about whether a topic can be taught.

archived-teaching-research.json preserves two historical records excluded from active coverage after a topic merge and a change in learning scope. Each entry includes the original topic and research record, its archive reason, and related current topic IDs. The archive also records the assessment prerequisite reference redirected after the merge.

The active teaching-research.json file has schemaVersion: 1 and taxonomy version: "v1". Each records[] entry joins topics.json through topicId and contains:

Field	Meaning
pedagogy	Original teaching summary, practices, representations and misconceptions.
learn	Portable learning guidance (schemaVersion: 2): a summary, activities, educational constraints and review notes.
learn.activities[]	Learner actions, intended insight, feedback, rationale and supporting source claims. No product artifact types or fit rankings.
assess	Original diagnostic/task, response levels, error patterns and next teaching moves; nullable in the public interface.
sources	Claim, named source, optional URL and the original author's verification flag.
confidence	Original authoring confidence; not an independent quality score.
provenance	Original authoring model and source record timestamps. The update timestamp also reflects the learning-guidance rewrite.
publicationNotes	Paths and curriculum codes for text omitted under the codes-only publication policy.

sourceClaims exactly match sources[].claim after publication transformations. A source may support a general principle or curriculum alignment; this does not mean the proposed activity was tested. Some activities have no directly linked source claim.

These records were agent-authored. The learning guidance was rewritten for portability, checked by a separate factual critic, and sampled by a second agent. Most original pedagogy and assessment were retained and may contain claims that the updated guidance qualifies; chart-specific alignment repairs after #30 are recorded in the affected learning-guidance review notes; read the constraints and review notes alongside them. Sources were not re-fetched or re-verified for the initial publication; later nutrition-guide checks are recorded in the affected review notes. Historical verified: true and confidence: "high" are not new evidence ratings. Activities are research-informed proposals, not empirically validated interventions.

Nutrition terminology corrections in #63 update the existing Animal Nutrition, Balanced Diet & Food Groups, and Nutrients in a Healthy Diet guidance, activities and assessment to distinguish nutrients from food groups. Balanced-plate activities use food groups in a local guide; the records retain their identities. See the review note.

Some original research quoted restricted curriculum standards. Matched passages are replaced with inline curriculum references, including inside citation claims; all transformed paths are listed in publicationNotes. This is a conservative mechanical omission pass, not comprehensive rights clearance. Consult the linked upstream sources for their wording and terms. See PROVENANCE.md.

After the chart split in #30, mt_bESTSBB0wK covers pie charts, mt_H6LlpWgEYS covers earlier line-graph problem-solving, and mt_HlNSurPVDp has a separate construction-and-interpretation record. The new record adapts existing graph research; it is not a fresh literature review, and all reused source-verification flags are false. Its assessment requires a complete constructed graph. The later grouped-charts assessment routes graph-choice errors to data presentation and sector-angle errors to pie charts.

The file is bound to the exact topics.json bytes through topicFileSha256. Regenerate the research projection when the canonical topic file changes, even when topic IDs remain stable.

Using it

Pure data — no runtime, no dependencies. Load the JSON and go.

import topics from './data/topics.json' with { type: 'json' };
import deps from './data/dependencies.json' with { type: 'json' };

const byId = new Map(topics.topics.map(t => [t.id, t]));
const prereqs = deps.dependencies
  .filter(d => d.topicId === 'mt_N8CpN1EJrP')
  .map(d => byId.get(d.prerequisiteId).name);

Join teaching guidance to a topic:

import research from './data/teaching-research.json' with { type: 'json' };

const guidanceByTopic = new Map(research.records.map(r => [r.topicId, r]));
const guidance = guidanceByTopic.get('mt_N8CpN1EJrP'); // undefined: English is not covered yet
const activities = guidance?.learn.activities ?? [];

Validate structure, research coverage, evidence links and checksums:

node scripts/validate.mjs
node --test scripts/teaching-research.test.mjs

Marble Builder

The marble-builder agent skill helps turn a child's age and a user-confirmed goal into a standalone HTML learning experience that works offline by default. It uses topic prerequisites and available teaching research, develops a targeted research brief when needed, and agrees a concept with you before building. Experiences use visual interactions, scaffolding, UI-assessed transfer tasks and optional extensions. Spatial topics use Three.js with direct scene manipulation and a crafted cel-shaded look with flat colors and black outlines. Visual simulations must meet a beauty, craft and polish bar as well as work correctly. The public skill includes Marble-derived UI tokens, tactile controls refined in Moon Visit and Sky Colour Explorer, and the openly licensed Baloo 2 font, with one title and one instruction visible at a time: a small top-left title and a large, prominent instruction or explanation, plus concise button labels paired with icons. Water Lab is the visual quality reference. The agent renders, inspects and iterates in the browser rather than stopping at the first working HTML file.

The skill includes five runnable examples with readable source, key-state screenshots and design notes: Moon Visit, Sky Colour Explorer, Hidden Cubes, Bee Builder and Otter Forest. The notes explain what to borrow and where each prototype has limits or exceptions.

Optional connected LLM mode requires explicit author approval and a learning goal whose open-ended reasoning or text production cannot be adequately supported locally. That mode requires internet and a separate account connection; ChatGPT plan usage is a preview option for eligible local/open-source integrations.

In Codex, open this repository and invoke $marble-builder, for example:

Use $marble-builder to help my 8-year-old understand why equivalent fractions represent the same quantity.

Other agents can follow the linked SKILL.md directly. No Marble account or hosting is required. New reusable teaching research can optionally be proposed as a separate PR, with no child-specific information included.

License

This dataset is multi-licensed — read this before you use or redistribute it.

Layer	License
The database — the collection, structure, IDs, topic↔topic and topic↔standard relationships	ODbL 1.0 — free for research and commercial use, attribution required, share-alike (derivative databases must stay open under ODbL).
The textual content Marble authored — topic description/name/evidence/assessmentPrompt, dependency reasons, cluster summarys, original teaching-research prose	CC BY-SA 4.0 — same spirit: attribution + share-alike.
Marble-authored learning examples — application source, original artwork, captures and design notes	CC BY-SA 4.0. Bundled Three.js and Baloo 2 retain their separate MIT and OFL licenses and notices.
curriculum-standards.json — extracted from third-party frameworks	Not Marble's to relicense. Each source is under its own upstream license — see PROVENANCE.md.

Why share-alike + still commercial-friendly: ODbL distinguishes a derivative database (extend/modify the taxonomy → must stay open) from a produced work (use it inside a product, model, or app → stays yours). So you can build a commercial product on this without open-sourcing your product; you only owe back improvements to the taxonomy itself.

Attribution

Any use must credit:

Marble Skill Taxonomy (v1) · © Generative Spark, Inc. (Marble) · https://withmarble.com · licensed under ODbL 1.0 (database) and CC BY-SA 4.0 (content).

Plus the upstream notices in PROVENANCE.md for any curriculum standards you use. See CITATION.cff for a formal citation.

What's not here

Deliberately excluded from this release: semantic embeddings (derived, recomputable) and any per-child / user data (never published). See CHANGELOG.md.