sophie

Design or restructure a course or module with the Sophie course-design (Mode-A) skill — audit-first backward design, emitting a course-spec + LO-graph + prose outline.

Brainstorm how to teach a topic with the Sophie design partner — a patient, generative, Socratic pedagogy collaborator that explores ideas before committing and converges only when you say you're ready.

Draft a practice set for a section from a Sophie practice-set packet, then self-review and present a reviewable diff.

Draft or revise a reading for a unit from a Sophie reading packet, then self-review and present a reviewable draft.

Review a unit's reading against the Sophie evidence-based pedagogy rubric, separating the deterministic floor from quality-ceiling judgments.

Independent pedagogy reviewer for Sophie-authored science content. Use to judge a reading or a drafted practice set against the Sophie evidence-based pedagogy rubric and return severity-tagged findings. Dispatched by the practice-set author for an independent rubric pass, or to back /sophie:review-chapter.

Use when introducing an abstract concept and you want students to build intuition — grounds the abstraction in multiple, varied, structurally-matched real-world examples (not one, not surface-similar) and fades from concrete to abstract so the idea transfers. The fifth of the six evidence-based learning strategies.

Use when designing or restructuring a course or module — the strategic Mode-A front of the authoring funnel. Drives backward design as a ONE-move-at-a-time conversation — audit the existing course first (brownfield), derive outcomes → evidence → content, then emit a course-spec.yaml + pedagogy/lo-graph.yaml (validated by LoGraphSchema) + a prose outline. Holds the line under "just give me the lecture list" pressure by naming the content-first tradeoff instead of dumping a curriculum. Composes evidence-based-pedagogy; hands LOs to the reading/practice authors. Platform skill; the course adds voice + domain data.

Use when a concept is served by a diagram and you're pairing a visual with prose — pairs COMPLEMENTARY verbal + visual representations of the same idea (Paivio's two channels; Mayer's multimedia principles — contiguity, coherence, redundancy-avoidance, signaling), not decorative or redundant images. Composes figure-authoring for figure quality. NOT learning styles.

Use when authoring explanations, reveals, or reflection prompts — prompts students to generate the "why" themselves (self-explanation, elaborative interrogation) instead of writing the reasoning for them, and upgrades predict-then-check into predict-then-explain.

Use when presenting a governing equation or key relationship in a reading or lesson — enforces unpacking every symbol with BOTH meaning AND units, an explicit dimensional check, connecting the equation to an observable, and reading its structure (exponents/scaling); decides <KeyEquation> registry vs inline math and defers mechanics to the component reference.

1 hook across 1 event