Skip to content
Data
Skill

/latex-posters

OpenRouter API key for optional schematic generation and review.

From plugin
scientific-agent-skills
56k177 skills

Claude Code

$ npx -y skills add K-Dense-AI/scientific-agent-skills --skill latex-posters --agent claude-code

SKILL.md

latex-posters.SKILL.md
name: latex-posters
description: "Creates research posters in LaTeX using beamerposter, tikzposter, or baposter. Use for conference posters, academic presentations, multi-column scientific layouts, figure integration, typography, compilation, and PDF preflight."
allowed-tools: Read Write Edit Bash
compatibility: Requires a LaTeX distribution with the selected poster class and Poppler for PDF checks. Optional AI schematics require Python 3.10+, requests, network access, and OPENROUTER_API_KEY. baposter needs a separately supplied class file.
metadata:
  version: "1.10"
  last-reviewed: "2026-09-30"
  skill-author: K-Dense Inc.
  openclaw:
    primaryEnv: OPENROUTER_API_KEY
    envVars:
    - name: OPENROUTER_API_KEY
      required: false
      description: OpenRouter API key for optional schematic generation and review.

LaTeX Research Posters

Create a conference-sized PDF from scientific content, editable LaTeX, and verified figures. Use for new posters, paper-to-poster adaptation, institutional templates, and poster compilation or print preflight.

Choose the poster package

| Package | Use when | Reviewed target | | --- | --- | --- | | beamerposter | Beamer blocks or institutional Beamer themes fit the project | CTAN 1.13 | | tikzposter | TikZ styling and automatically stacked blocks are useful | CTAN 2.0 | | baposter | A supplied project uses its named, relative-positioned boxes | Archived class v2.0, 2011/11/26; upstream site unavailable at review |

The three bundled templates were compiled locally with pdfLaTeX (TeX Live 2025) and inspected using Poppler 26.09.0. baposter was tested using the pinned archival source described in [references/latex_poster_packages.md](references/latex_poster_packages.md). Do not interpret an old stable package version as proof of current tagged-PDF support. Other reference snippets are illustrative building blocks unless explicitly tested; they require the indicated class, packages, actual data, and figure files.

Workflow

1. Establish the scientific content and output requirements

  • Record exact paper width, height, orientation, venue rules, and printer requirements.
  • Choose 1-3 main messages and the results needed to support them. A useful starting

budget is 300-800 words with 3-5 figures; adjust to the audience and available space.

  • Preserve study design, sample sizes, units, effect sizes, uncertainty, limitations,

and data provenance. Include the relevant citations, contact, and acknowledgments.

  • Use actual analysis outputs for plots and original research images for evidence.

Do not use an image model to invent numerical charts, microscopy, spectra, or results.

2. Prepare and review figures

Prefer vector PDF for numerical plots, TikZ for precise diagrams, and suitable PNG/JPEG for photographs or conceptual artwork. Convert SVG to PDF before ordinary `\includegraphics`; SVG is not a directly supported pdfLaTeX image format.

Optional AI schematics can help explain a concept. Generate them only when useful; there is no required AI-image quota. Keep titles, citations, numerical values, and QR codes editable in LaTeX or a deterministic plotting tool. Use official logo assets.

# From the skill root; save project output outside the installed skill directory.
# Illustrative paid call, not executed during this review.
python scripts/generate_schematic.py \
  "Conceptual A0 poster diagram: three boxes SAMPLE, MEASURE, ANALYZE connected left to right. Large dark labels on white, generous margins, no numerical results." \
  -o /path/to/project/figures/methods.png --doc-type poster --iterations 1

The helper uses OpenRouter's `POST /api/v1/images` with Nano Banana 2 (`google/gemini-3.1-flash-image`) and reviews through `POST /api/v1/chat/completions` with `google/gemini-3.7-flash`. It accepts PNG output paths and 1-2 iterations. Generation and review both consume API credits. A saved image or an AI score does not establish scientific correctness or final print quality. Check the image and `*_review_log.json`, including `quality_met` and `termination_reason`.

Inspect every figure before assembly: relationships, labels, units, colors, license, and accuracy. Check actual placed size, pixel dimensions, and raster PPI. A requested โ€œ120 ptโ€ font in an image prompt is not a physical typesetting guarantee. See [references/ai_graphics_for_posters.md](references/ai_graphics_for_posters.md).

3. Create the layout

Copy one of the three files in `assets/` into the project:

  • `beamerposter_template.tex`
  • `tikzposter_template.tex`
  • `baposter_template.tex`

Each template compiles with visibly marked draft figure boxes when its assets are missing. These are placeholders, not example research findings. Replace them, example references, author details, logo boxes, and the example QR target before delivery.

Use 2-3 columns as a starting layout. Keep sufficient page margins and gaps between blocks. Figure widths must fit the containing box; `0.85\linewidth` is a reasonable starting point, not a universal maximum. Body text around 24-36 pt at final print size is a starting range. Font commands vary with class and scale; verify the actual rendered sizes rather than assuming that `\Large` is always a fixed size.

Package details, valid option keys, and layout caveats: [references/latex_poster_packages.md](references/latex_poster_packages.md).

4. Compile and inspect the actual PDF

Run these from the poster project directory:

pdflatex -interaction=nonstopmode -halt-on-error poster.tex
pdflatex -interaction=nonstopmode -halt-on-error poster.tex
pdfinfo poster.pdf
pdffonts poster.pdf
pdfimages -list poster.pdf
pdftoppm -scale-to 2000 -singlefile -png poster.pdf poster-preview
rg -n 'Overfull|Underfull|Warning|undefined' poster.log

Use LuaLaTeX/XeLaTeX if a project needs `fontspec` and installed OpenType fonts; changing engines still requires compilation and visual verification. Run BibTeX or Biber only when

Read more
@k-dense-ai
Ships withscientific-agent-skills

๐Ÿ”” Claude Scientific Skills is now Scientific Agent Skills. Same skills, broader compatibility โ€” now works with any AI agent that supports the open Agent Skills standard, not just Claude.

Stars on GitHub

Momentum

stars in total
55.6k
Star history of K-Dense-AI/scientific-agent-skills on GitHub, a chart by star-history.com

Other skills on scientific-agent-skills.

aeon

aeon

This skill should be used for time series machine learning tasks including classification, regression, clustering, forecasting,โ€ฆ

More skills from other plugins.

paddleocr-doc-parsing

paddleocr-doc-parsing

Use this skill to extract structured Markdown/JSON from PDFs and document imagesโ€”tables with cell-level precision, formulas asโ€ฆ

@paddlepaddle@paddlepaddleInstall Skill
context-loader
Auto-invoked

context-loader

Search memories from earlier Pi sessions in this repository. Use it when earlier work may already explain the code, error,โ€ฆ

mem04d69k
@mem0ai@mem0aiInstall Skill
analyzing-ai-citations

analyzing-ai-citations

Answer AI citation and answer engine optimization (AEO) questions in PostHog: whether AI answer engines (ChatGPT, Claude, and anโ€ฆ

posthog10d41k
@posthog@posthogInstall Skill
qmd

qmd

Search local markdown knowledge bases, notes, docs, and wikis with QMD. Use when users ask to find notes, retrieve documents,โ€ฆ

qmd3w31k
@tobi@tobiInstall Skill

Questions about latex-posters.

Which coding agents can use latex-posters?
latex-posters installs in Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, OpenCode and Windsurf with the skills CLI, using each agent's flag shown on this page. Flowy's auto-invocation, which fires a skill when your prompt matches, runs in Claude Code only.