CommunityImagengithub.com

scientific-visual-explainer

Build biomedical mechanism schematics, graphical abstracts, experimental workflow panels, and interactive or animated scientific lessons (PNG, editable SVG sources, requested PPTX, offline HTML, and Manim video). Use for Figure 1A schematics, pathway models, mechanism diagrams, pathway diagrams, graphical abstracts, and experimental flowcharts, redrawing static mechanism figures into editable components, or teaching papers/concepts visually. Default new schematics to Nord mist, capsule proteins and Y/U receptors; use domain-bearing structural detail when requested.

¿Qué es scientific-visual-explainer?

scientific-visual-explainer is a Claude Code agent skill that build biomedical mechanism schematics, graphical abstracts, experimental workflow panels, and interactive or animated scientific lessons (PNG, editable SVG sources, requested PPTX, offline HTML, and Manim video). Use for Figure 1A schematics, pathway models, mechanism diagrams, pathway diagrams, graphical abstracts, and experimental flowcharts, redrawing static mechanism figures into editable components, or teaching papers/concepts visually. Default new schematics to Nord mist, capsule proteins and Y/U receptors; use domain-bearing structural detail when requested.

Compatible con~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/mianaz/biomed-skills/tree/HEAD/skills/scientific-visual-explainer

Installed? Explore more Imagen skills: steipete/songsee, affaan-m/frontend-design, mattpocock/improve-codebase-architecture · View all 6 →

Preguntar en tu IA favorita

Abre un nuevo chat con esta habilidad de agente ya precargada.

Documentación

Scientific Visual Explainer

Read shared principles.

Make biological mechanisms, experimental workflows and scientific models visible. Read visual standards for shared Arial typography, nomenclature and independent panel rules.

Choose the deliverable route

First choose the explanatory purpose: graphical abstract, review schematic, mechanism model or visual explainer. Then select the output mode from the user's request before planning scenes:

The table gives defaults for each mode; a narrower requested format takes priority. For planning-only work, deliver the requested storyboard or plan and stop there.

RequestDeliverable routePrimary references
Static manuscript schematic, graphical abstract, or workflow panel (e.g., for Figure 1A, pathway summary, or figure-plan assembly)600 dpi PNG with a reusable editable SVG/scene source; add PDF, delivered SVG or native PPTX when requested or needed for assembly. Export without a publication panel letterbasic-mode.md, composition.md
Redraw or animate an existing static mechanism figure / graphical abstractSource-preserving component extraction (scene.json), original-vs-reconstruction comparison, editable PPTX, offline HTML and optional Manim MP4figure-to-scene.md
Interactive concept, review or paper explanation (or when no format is specified for a teaching request)Seekable offline HTML lesson (renderAt(time)) coupling the complete static mechanism figure with step-by-step animated events and condition comparisonsbasic-mode.md, visual-language.md, composition.md
Domain-detailed molecular interaction or animated videoReusable SVG icon rig from Biomed Icons (or generated/vectorized via generated-icons.md) with fitted contact contours, plus HTML/PPTX/Manim exportillustrations.md, generated-icons.md, composition.md, media.md

1. Establish the scientific model and evidence scope

For a graphical abstract, read graphical abstract to select and compress the contribution before choosing assets.

For a pathway database or predefined pathway layout, read pathway-assembly.md to import identities and geometry, then map them to reviewed mechanism events and biological anchors.

Read the supplied manuscript, paper, review or figure first:

  • Keep measured results, proposed mechanisms and illustrative teaching models distinct. Link each consequential claim to an exact panel, section or page.
  • Preserve literal biological entities, compartments, membrane topology, strand direction and species nomenclature. Identify whether each arrow represents transport, biochemical reaction, measured effect or proposed regulation (activation arrow vs. inhibition T-bar).
  • Save storyboard.md recording the target audience, visual goal, entity/compartment inventory, and scene or panel table (question / entities / visual change / takeaway / source).

Done: storyboard.md defines the exact entities, compartments, arrow meanings and source evidence boundaries.

2. Select symbol mode and rig assets

Search the unified icon registry before drawing new assets. Reuse the same asset key for the same subject across figures; choose its recorded view, state and palette. Local reference redraws join the existing catalog without replacing accepted geometry or creating a separate scene library.

Inspect relevant BioRender public examples and published CNS mechanism figures (from local PDFs or curated galleries) to choose clear compartment layouts and symbol flow; record visual_references separately from scientific_references.

  • Basic symbolic mode (default for new schematics, graphical abstracts and lessons): Read basic-mode.md and start from the 38-symbol primitive library and three basic mechanism examples.
    • Use Nord mist (assets/basic-mechanisms/palettes.json), generic-protein-capsule for proteins, generic-receptor-y / generic-receptor-cup (and their dimers) for receptors, adjacent capsules for complexes, and separate P circles for phosphorylation.
  • Detailed structural mode (when requested or required by domain-specific mechanism): Read illustrations.md and reuse Biomed Icons first (egfr-imagegen, tgfb-imagegen).
    • When a new structural icon is needed, follow generated-icons.md to generate, vectorize, curate parts/anchors and register it.
    • When molecules physically dock, read the contact-contour rules in composition.md so binding partners share a smooth closed interface at assembly scale.
  • Existing static figure import: Follow figure-to-scene.md (scripts/figure_to_scene.py, scripts/build_figure_lesson.cjs, scripts/export_figure_pptx.mjs) to recover grouped SVG/raster components into scene.json.

Done: all required icons, compartments and anchors are selected or registered with a contact sheet and provenance record.

3. Build the schematic or animated lesson

Compose a clean static view at the selected explanatory level before adding animation beats. A graphical abstract uses the content selected above. Apply the artwork title rule to GA/review schematics and shared source/export policy to reusable sources and requested formats:

  • Keep labels as separate editable SVG/PPTX text in Arial. Put protein names inside their capsules and size the capsule to fit. Keep compartment and complex labels outside. Check the text against the capsule boundary and adjacent bodies.
  • Fit a ligand's contact contour to the corresponding receptor opening. Preserve the accepted receptor geometry and check the mating curves, contact clearance, and fill boundaries in both free and bound states.
  • For static manuscript panels / graphical abstracts, export a letter-free 600 dpi PNG at its intended placed size, retaining the editable scene/SVG source. Use export_figure_pptx.mjs or a native shape builder when editable PowerPoint assembly is requested; generate other formats only for the requested delivery.
  • For interactive HTML lessons, drive drawing, captions and condition comparisons from a single state via renderAt(time) using scripts/domain-motion.js, scripts/mechanism-arrows.js and scripts/mechanism-annotations.js. Keep the complete static overview one click away. Reference worked lessons: EGF–ERK, TGFβ–SMAD, and Cas9.
  • For slides or Manim video, follow media.md and scripts/render_figure_manim.py.

Done: the static schematic and any requested HTML/PPTX/video outputs are built from a single consistent scene model.

4. Inspect at viewing size and deliver

Open every static frame, PPTX slide, or interactive scene in its intended use: graphical-abstract thumbnail, manuscript column, presentation or interactive viewport. For responsive lessons, also inspect a narrow viewport. Reflow detail for that use while preserving the model and the shared effective-size rules:

  • Check for text/text, text/icon, arrow/icon and phospho-badge collisions across initial, intermediate approach, bound and released states.
  • Verify compartment topology (receptors span the membrane midplane; nuclear import crosses a visible pore; cargo stays on the correct lumenal/cytosolic side).
  • Verify that displayed quantitative readouts match source data or are explicitly labeled illustrative / simulation / digitized.

Apply reader read-back to the final view and caption as part of this inspection.

Done: visual defects are resolved on the rendered artifacts. Deliver the requested formats and link the reusable source, storyboard and model. A static PNG delivery does not require extra PPTX/PDF/HTML exports.

Individual skills in this repo

This repo contains 1 individual skill — each has its own dedicated page.

Skills relacionados