Research Poster

---
name: research-poster
description: >
  Turns a PDF abstract into a polished, fully-editable scientific conference poster (.pptx + PDF +
  preview) in a matched house style, with honest charts, verified citations, big readable
  typography, and a render-and-QA loop. Use whenever the user wants to make, build, design, or
  create a research poster, scientific poster, academic poster, conference poster, or ePoster from
  an abstract, paper, or study; to imitate an existing poster's style with new content; or when they
  mention turning an abstract or paper into a poster, a poster from a PDF, or updating a poster's
  data, charts, layout, or making poster text bigger or cleaner. Builds native, editable PowerPoint
  charts by default (avoiding the PowerPoint chart-repair prompt), or polished image charts on request.
---

<objective>
Produce a submission-ready, on-brand scientific poster from a PDF abstract. The deliverable is an
editable PowerPoint (.pptx) plus a PDF and a preview PNG. Everything is native and editable by
default: text, section bands, stat tiles, and charts. The data, numbers, and citations are always
real and verified; the design is readable from a distance, breathable, and matched to the author's
house style (extracted from a reference poster when one is provided).

This skill encodes a workflow that was hardened in practice. Read the bundled references before the
step that needs them; do not reinvent the parts that are already solved (chart-safety, rendering,
typography scale). Scripts live in `${CLAUDE_SKILL_DIR}/scripts/`; worked examples and deep guidance
live in `${CLAUDE_SKILL_DIR}/references/`. Rendering uses LibreOffice (cross-platform) or Microsoft
PowerPoint on macOS; see `references/rendering-qa.md`.
</objective>

<requirements>
- python 3 with: python-pptx (>= 1.0), matplotlib, pdfplumber, pypdfium2, pillow, pypdf.
- A renderer: LibreOffice (`soffice`/`libreoffice` on PATH, any OS) OR macOS with Microsoft PowerPoint.
  The repair-detector (`scripts/detect_repair.sh`) is macOS + PowerPoint only and is optional.
</requirements>

<execution_context>
Run interactively OR fully autonomously. Every gate below is an `AskUserQuestion` with a clearly
marked recommended default, so an autonomous run just takes the defaults and states them.

- 🟢 Auto (no ask): read inputs, research, extract house style, build, render, self-QA and iterate,
  git-init + commit in the poster's own working dir, clean up intermediates.
- 🟡 Ask (with a recommended default): poster dimensions / venue spec, author + affiliation + logo
  confirmation, tone level, citation depth, chart engine (editable-native vs images), and whether
  the data is final.
- 🔴 Never: fabricate any number, result, or citation; alter the study's reported findings; overwrite
  a file the user has already finalized (always write a separate copy); ship a file that triggers
  PowerPoint's repair prompt.
</execution_context>

<principles>
- Reiterate the plan and confirm before building when working with a user; pair questions with a
  recommendation so the workflow can also run autonomously.
- Explore and verify, do not assume. Research with web search and a web reader; consult current
  library documentation for any API before using it. Verify every citation against PubMed or a DOI.
- Keep it lean and DRY. Reuse the bundled scripts and examples; adapt, do not rebuild. Drive every
  chart and callout from one editable data block so numbers update in one place.
- Design deliberately: strong visual hierarchy, generous padding, high contrast, readable-from-a-
  distance type. Avoid neon, heavy gradients, glow, and AI-generated imagery. Match the author's
  venue conventions.
- TDD-lite: assert every chart percentage recomputes to the source number before building.
- Self-QA to a high bar before showing the user: render, look, fix, repeat.
- Never fabricate data or citations, and never alter reported findings.
</principles>

<process>

<phase name="0_intake_and_confirm">
## Phase 0: Intake + confirm

1. Collect inputs: the abstract PDF (required); an optional reference poster (.pptx) for house
   style; optional logo files; any updated data.
2. Reiterate what you understood, then batch the gates (one or two `AskUserQuestion` calls, each
   option marked with the recommended default). Recommended defaults in brackets:
   - Dimensions / venue: verify the venue's ePoster spec first [default]; else 16:9 single slide
     (48x27 in) which matches most on-screen ePoster viewers.
   - Authors, affiliations, and logos: confirm against the author, not the abstract's submission
     block (it is often not final). Multiple institutions means multiple logos [ask which].
   - Tone: moderate scholarly polish [default]; data and findings never change.
   - Citations: 4-6 verified references with clickable DOIs/URLs [default].
   - Chart engine: editable native charts [default]; polished images on request.
   - Is the data final, or is more coming? If ongoing, build so numbers are trivial to update
     (single data block) and keep the preliminary framing.
3. Autonomous mode: take the defaults, state them, proceed.
</phase>

<phase name="1_read_inputs">
## Phase 1: Read every input fully

1. Optionally use companion skills if the environment has them (document/pptx, dataviz, frontend
   design, scientific writing, citation management, scientific visualization). They are enhancements;
   the workflow below is self-contained via python-pptx, matplotlib, and a renderer.
2. Read the abstract completely: text via `pdfplumber`, a visual render, and an OCR/image check so
   no figure is missed. Capture the title, authors, affiliations, every number, and the n/N for each.
3. If a reference poster is provided, extract its house style: slide size, palette (band + accent
   fills), fonts and sizes, section-bar structure, and the logo. Render it with
   `bash ${CLAUDE_SKILL_DIR}/scripts/render.sh` to see it. See `references/design-rules.md`.
4. Confirm python deps: `pip install python-pptx matplotlib pdfplumber pypdfium2 pillow pypdf`.
</phase>

<phase name="2_research">
## Phase 2: Research

1. Verify the venue's ePoster spec (dimensions, format, template).
2. Verify and fetch citations: real, PubMed/DOI-checked, with hyperlinks and access dates; include
   registry and grey-literature sources the author actually uses, not only journals. See
   `references/citations.md`. Never invent a citation.
3. Pull any background needed for the introduction. Consult current library docs when building.
</phase>

<phase name="3_design">
## Phase 3: Design

1. Set house-style tokens, a column grid, clear visual hierarchy, poster-scale typography, breathable
   padding, and strong contrast. See `references/design-rules.md`.
2. Choose chart types honestly: doughnut/pie only for parts-of-a-whole; a bar chart for independent
   or non-mutually-exclusive proportions; big-number tiles for single figures. Label every chart
   with its n/N. See `references/build-and-charts.md`.
3. Define one editable data block that drives every chart and callout.
</phase>

<phase name="4_build">
## Phase 4: Build (python-pptx, TDD-lite)

1. Adapt `references/examples/build_poster_example.py` (editable native charts) or
   `references/examples/make_charts_example.py` (matplotlib images) to the content and house style.
2. Build native editable charts with PowerPoint-safe styling ONLY. The single hard rule:
   never set a doughnut/pie data-label POSITION, that is the setting that makes PowerPoint
   repair-and-remove the chart. Per-slice fills, legends, percentage labels, number formats, and
   axis fonts are all safe. Put the doughnut percentage in the legend plus a big number in the hole.
   Full recipe in `references/build-and-charts.md`.
3. TDD-lite: assert every displayed percentage recomputes to the source number before saving.
4. Get authors, affiliations, and logos right per Phase 0.
</phase>

<phase name="5_render_and_self_qa">
## Phase 5: Render + self-QA loop

1. Render with `bash ${CLAUDE_SKILL_DIR}/scripts/render.sh <pptx> <out.pdf> <out.png> 1.7`
   (LibreOffice, or PowerPoint on macOS), then view the PNG.
2. On macOS with PowerPoint, optionally run `bash ${CLAUDE_SKILL_DIR}/scripts/detect_repair.sh <pptx>`
   to confirm native charts survive with no repair alert.
3. Fix against the `references/rendering-qa.md` checklist: no overflow or clipping, big enough type,
   breathable padding, high contrast, aligned columns, correct numbers. Iterate until it is genuinely
   good before showing the user.
</phase>

<phase name="6_deliver_git_cleanup">
## Phase 6: Deliver + git + cleanup

1. Write a SEPARATE, clearly named copy; never overwrite a file the user has finalized.
2. Deliver the .pptx, a submission PDF, and a preview PNG.
3. git-init the poster's working dir if needed and commit with a clear message (baseline first when
   iterating on an existing poster, so the prior version is preserved).
4. Remove intermediate renders and temp files; keep the reproducible build scripts.
</phase>

</process>

<success_criteria>
- [ ] Deliverable is an editable .pptx plus a matching PDF (verified page size) and preview PNG
- [ ] Charts are native and editable (default) and survive PowerPoint open with no repair prompt
- [ ] Every percentage recomputes to the source number; n/N shown; preliminary framing kept if ongoing
- [ ] Citations are all real and verified, with DOI/URL hyperlinks
- [ ] Authors, affiliations, and logos confirmed with the user
- [ ] Type is readable from a distance, padding is breathable, contrast is strong, columns align
- [ ] House style matches the reference poster when one was provided
- [ ] A separate copy was written; the user's finalized files were not touched; work is git-tracked
</success_criteria>

<error_handling>
- render.sh prints "RENDER FAILED" if no renderer is found: install LibreOffice, or run on macOS with
  Microsoft PowerPoint. It already retries the PowerPoint path once on a transient launch race.
- A native chart went missing after open (detect_repair.sh reports fewer charts, macOS only): a
  doughnut/pie data-label position was set somewhere; remove it (see references/build-and-charts.md).
- python dep missing: `pip install python-pptx matplotlib pdfplumber pypdfium2 pillow pypdf`.
</error_handling>

<references>
Read on demand, one level deep from here:
- `references/build-and-charts.md` : python-pptx build approach, the native-editable-chart safe recipe
  and the exact repair trigger to avoid, honest chart-type choices, and chart styling.
- `references/design-rules.md` : poster-scale typography and hierarchy, padding, contrast, house-style
  extraction from a reference .pptx, logos, and confirming the author block.
- `references/rendering-qa.md` : the cross-platform render path, robustness, the repair detector,
  and the self-QA checklist.
- `references/citations.md` : verifying, formatting (AMA/Vancouver), hyperlinking, and sourcing.
- `references/examples/` : two worked build scripts (native charts, image charts) to adapt.
</references>