Figure Composer — narrative → panels → compose → adversarial loop

figure-composer is the outer workflow for one multi-panel figure. Use the figure-style rules while planning and reviewing; every panel worker uses those rules independently. Run paper-narrative first when the paper-level figure sequence is still undecided.

Open-Science Notebook call

Every notebook_execute request whose code uses a function named in this skill includes this skill ID:

{ "kernelSkillIds": ["figure-composer"], "code": "print(figure_outline_schema())" }

kernelSkillIds contains the skill ID; function calls belong in code. Call the named functions directly without an import or discovery step.

Inputs

• claim: the one sentence the figure makes true without surrounding prose.
• dataVersionIds: immutable Upload or Artifact Version identities grounding the panels.
• width_mm: venue column width, commonly 85–89 mm single or 174–183 mm double.
• rulesVersionId: immutable Artifact Version containing the design rules used by the composite reviewer.
• delegatePrefix: short branch-unique prefix for panel and reviewer child names.

Run this workflow only in the Main/root agent. Delegated children cannot call host.delegate, so the whole composer cannot itself be delegated.

Entry points

• From a claim: Main writes the outline in step 1 from the claim, data, and figure-style rules.
• From an existing figure: inspect it with host.viewImage, then have Main draft and review the outline directly. Current host.llm calls do not accept images, so do not add a second hidden inference step. Pixels cannot supply Artifact Version identities; fill data_vid from the provided data.

1. Narrative → panel outline

Main produces a panel_outline matching figure_outline_schema():

{
  "claim": "…",
  "width_mm": 180,
  "ncol": 12,
  "row_heights_mm": [40, 60, 46, 52],
  "panels": [
    {
      "letter": "a",
      "role": "schematic",
      "row": 0,
      "col": 0,
      "colspan": 12,
      "chart_family": "schematic overview",
      "message": "…",
      "data_vid": null,
      "ask": "…"
    },
    {
      "letter": "b",
      "role": "primary",
      "row": 1,
      "col": 0,
      "colspan": 7,
      "chart_family": "scatter + trend",
      "message": "…",
      "data_vid": "…",
      "ask": "…"
    }
  ]
}

Outline rules:

• A is the context-free hook: schematic or hero, normally full width.
• B carries the claim: it should make the sentence true on its own.
• Remaining panels add evidence in descending importance.
• Use one row per sub-claim, normally 5–10 panels, and a 12-column grid.
• Every non-schematic data_vid must be one of the supplied immutable Version identities. Do not invent or rewrite Version IDs.
• Set fixed_panel_set: true only when the user explicitly requires the exact listed panels.

Geometry helpers reject duplicate panel letters (case-insensitive), overlapping grid spans, panels outside the grid, and invalid or subpixel grid dimensions. Use unique panel identifiers and non-overlapping positive spans within the grid.

Review the outline before fan-out. Use the schema as a contract; Main does the reasoning and does not call host.llm to generate the outline again.

2. Fan out panel workers

Generate each task in Python with panel_task(outline, letter, fig_label). The returned task contains the complete panel procedure. Pass it unchanged on the first render and supply the panel's data Version in inputs.

Dispatch from repl_execute. host.delegate accepts at most four children per atomic call, so send ordered waves of no more than four. Each request uses this output schema:

const panelOutputSchema = {
  type: 'object',
  additionalProperties: false,
  required: ['panelVersionId', 'labelsUsed'],
  properties: {
    panelVersionId: { type: 'string', minLength: 1 },
    labelsUsed: { type: 'array', items: { type: 'string' } }
  }
}

Use wait: false, then collect the exact { frameId, attemptId } receipt handles. A collect timeout ends observation, not the child Attempt: collect the same handles again while any remain running. Retry only after a terminal failure or an explicitly rejected output, using a fresh child name. Panel workers must submit their structured result with host.submitOutput before finishing. Reject a non-completed/error child, missing or unsatisfied structured output, a missing or duplicate expected panel_<letter>.png, or a mismatch between its Artifact versionId and structuredOutput.panelVersionId. MIME metadata may be absent; the exact filename and Version identity are the binding checks. Return each wave's validated { letter, versionId } values from the repl_execute call instead of relying on local const or let declarations to survive a later call.

Keep finalized Version identities in outline order. Temporary paths are never the Agent-to-Agent contract. Child names remain occupied after settlement, so use a unique delegatePrefix and round number.

3. Compose and bind the producer Run

Generate a producer task with composition_task(outline, panelVersions, fig_label). Main's newly written Artifact can remain pending until its turn ends; the producer child publishes a finalized composite that the reviewer can use. Pass the ordered panel Version identities in inputs and require this output schema:

{
  type: 'object',
  additionalProperties: false,
  required: ['compositeVersionId'],
  properties: { compositeVersionId: { type: 'string', minLength: 1 } }
}

The producer resolves the collected Version identities and places the paths in a small JSON handoff under process.env.OPEN_SCIENCE_HANDOFF_DIR. On its notebook_execute request, it passes the ordered, de-duplicated panel identities as artifactVersionInputs. This registers the delegated immutable panel Versions as the composition Run's provenance inputs; paths remain byte-access implementation details and must never replace Version identities in this field. The producer calls compose_figure, verifies notebook completion, and keeps the actual returned runId. It publishes the final PNG with write_artifact_file({ filename: "figure.png", producerRunId: composeResult.runId }); never substitute a round number or locally invented Run identity. This binds the composite Artifact to the run that last wrote its bytes. Fail the workflow if any panel Version cannot be validated in the active Project; never silently compose with an unregistered provenance input.

Collect the exact producer Attempt and require completed status, satisfied structured output, and exactly one figure.png Artifact whose versionId matches structuredOutput.compositeVersionId. Use that finalized composite Version for inspection and review. The producer submits the structured result with host.submitOutput and finishes normally.

compose_figure requires each input image to match its panel_px dimensions exactly. A mismatch raises before the output is saved; regenerate the panel at the requested size. Images are never stretched to fit. Use the exact figsize expressions generated by panel_task, rather than rounded inch measurements, and verify the saved PNG dimensions.

3.5 Look before review

Call compose_crops in Python and inspect every crop before formal review. host.viewImage never upscales and caps the output long edge at 1568 pixels; omit maxSize when native pixels are required.

One repl_execute invocation can attach at most four images. Split five or more crops into ordered batches of no more than four, and let each invocation finish successfully before starting the next; a failed enclosing invocation discards every image staged by that invocation. For each cropBatch, use the current camelCase API:

if (cropBatch.length > 4) throw new Error('viewImage crop batch exceeds four images')
for (const [letter, box] of cropBatch) {
  await host.viewImage(
    { versionId: compositeVersionId },
    { crop: { unit: 'pixels', left: box[0], top: box[1], right: box[2], bottom: box[3] } }
  )
}
return { inspectedPanels: cropBatch.map(([letter]) => letter) }

Check contrast, smallest marks, leader crossings, color identity, legend binding, seams, panel-letter overlap, gutter bleed, and resize artifacts. Fix an obvious defect before formal review.

4. Adversarial review loop

Run at most three rounds. An independent reviewer Attempt is required before returning any composite. Generate the reviewer task with composite_review_task(...) and its outputSchema with review_schema(). Pass the task unchanged to one reviewer; include the composite, optional previous composite, rulesVersionId, and every non-null panel data Version in inputs. Collect the exact receipt and use only validated structuredOutput as the review object. The reviewer submits it with host.submitOutput; do not replace formal review with Main's own inspection.

After each result:

1. Accept when the verdict is accept or minor_revision, there are no BLOCKERs, and there are at most two MAJORs.
2. Save previous_outline = copy.deepcopy(outline) before applying outline_revisions explicitly. Then call apply_outline_revisions(outline, revisions, previous_outline=previous_outline). This includes new panels and every panel whose pixel dimensions changed, even when a shared row-height change names only one panel. Pass the same dpi and gutter_mm as composition if overriding their defaults. Removed panels are excluded; drop their entries from the collected panel Versions.
3. Call group_fixes_by_panel(review) and compute regen = (affected | set(fixb)) & {p["letter"] for p in outline["panels"]}.
4. Regenerate only regen. Build each retry task as panel_task(outline, letter, fig_label) + fixb.get(letter, "") and add: “Do not over-correct: preserve everything the previous version got right.” Include the prior panel Version when one exists and its data Version in inputs.
5. Keep every clean panel's exact Version identity. Compose a new revision with a fresh producer child only after every regenerated panel passes the same identity checks. Review only that new composite Version.

Stop when accepted, or when outline_revisions is empty and new findings are only carve-out exceptions to the previous round; that is the over-labeling signal. Otherwise stop after round three. If the current composite was not accepted, report the unresolved findings rather than return an older composite as the final result.

After acceptance, verify the composite's provenance contains the current panel Versions. Return that finalized figure.png Artifact with a user-visible link; do not publish a duplicate root Artifact.

Anti-patterns

• Do not regenerate clean panels.
• Do not manufacture findings.
• Verify review anchors on the composite, not only on isolated panels.
• Remove labels that a reader with field context would find redundant.