anti-ai-poster-design
A model-agnostic skill that helps any AI make posters that don't look AI-generated
Safety scan
A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.
Check it deeper on your machine
It has changed since its scan on 20 Nov 2026.
As a Claude Code plugin, install it and then ask Claude Code:
/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/sayan1997/anti-ai-poster-design with scanskillsafety
Or from a terminal, clone it and run it:
git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/sayan1997/anti-ai-poster-design'
It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.
Badge for your README
Markdown
[](/listings/sayan1997--anti-ai-poster-design)HTML
<a href="/listings/sayan1997--anti-ai-poster-design"><img src="/api/listings/sayan1997--anti-ai-poster-design/badge.svg" alt="findsafeskills safety rating" /></a>Install
What was scanned
The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.
README.md · README · 90 lines
- # Anti-AI Poster Design
- **A model-agnostic skill that helps any AI make posters that don't look AI-generated.**
- Most AI-made posters share a fingerprint: gold-gradient lettering, ornate filigree frames, a pill or ribbon behind every line, glowing glossy food, icon-in-circle grids, five fonts at once, garbled fine print. They fail because *nothing is ever removed*. This skill teaches a model to design the way a human does: **one idea, one hero, one action, then delete.**
- It was built from a side-by-side analysis of 12 AI-style street banners and 8 human-designed posters (see [`references/ai-tells-catalog.md`](skills/anti-ai-poster-design/references/ai-tells-catalog.md)), and includes Bengali / Indic typography guidance.
- ## What's inside
- | Path | What it is |
- |---|---|
- | [`skills/anti-ai-poster-design/SKILL.md`](skills/anti-ai-poster-design/SKILL.md) | The skill: workflow, hard limits, the list of tells |
- | [`references/ai-tells-catalog.md`](skills/anti-ai-poster-design/references/ai-tells-catalog.md) | Evidence: ~60 catalogued tells with IDs, measurements, replacements |
- | [`references/design-system.md`](skills/anti-ai-poster-design/references/design-system.md) | 5 layout skeletons, type scale, 10 palettes, font pairings, quality tests |
- | [`references/bengali-and-indic-typography.md`](skills/anti-ai-poster-design/references/bengali-and-indic-typography.md) | Rendering, sizing, bilingual layout, numerals, cultural notes |
- | [`references/image-model-prompting.md`](skills/anti-ai-poster-design/references/image-model-prompting.md) | Two-stage workflow and prompts for Midjourney, DALL-E, Stable Diffusion, etc. |
- | [`assets/template.html`](skills/anti-ai-poster-design/assets/template.html) | Starter poster with palette/type tokens |
- | [`scripts/`](skills/anti-ai-poster-design/scripts) | `render_poster.py` (HTML to PNG), `poster_lint.py` (source audit), `poster_pixels.py` (image audit) |
- | [`examples/`](skills/anti-ai-poster-design/examples) | Three worked posters (HTML + PNG) |
- | [`PORTABLE_PROMPT.md`](PORTABLE_PROMPT.md) | One paste-able file for any chatbot |
- ## Examples
- All copy in the examples is sample/placeholder content. Each was built with the skill and passes the linter.
- | Giant Word (Bengali) | Traditional Mode (Bengali) | Ledger (dark ground) |
- |---|---|---|
- |  |  |  |
- ## Use it
- ### Claude (claude.ai, Claude Desktop, Claude Code)
- 1. Download this repository as a ZIP, or download the `.skill` file from the Releases page.
- 2. **claude.ai / Claude Desktop:** upload the skill from your Claude skills settings.
- 3. **Claude Code:** copy the folder `skills/anti-ai-poster-design` into `~/.claude/skills/` (all projects) or `<your-project>/.claude/skills/` (one project).
- 4. Ask: *"Make a poster for my tea stall: kochuri and dalpuri, Bengali, for Instagram."* The skill triggers automatically.
- ### Any other AI (ChatGPT, Gemini, Copilot, local models, API)
- Open [`PORTABLE_PROMPT.md`](PORTABLE_PROMPT.md) and paste everything below the line into a system prompt, custom instructions, a Gem / custom GPT, a project, or the top of a chat. It is self-contained and needs no tools.
- For models that can browse a repository or read files, point them at `skills/anti-ai-poster-design/SKILL.md` and let them follow the links to the references.
- ### Image generators (Midjourney, DALL-E, Stable Diffusion, Imagen, Ideogram, Flux)
- Follow [`image-model-prompting.md`](skills/anti-ai-poster-design/references/image-model-prompting.md): generate the *hero image only, with no text*, then add the text in any layout tool or in HTML.
- ## Scripts
- ```bash
- # Render an HTML poster to PNG (Playwright+Chromium, or WeasyPrint+PyMuPDF as fallback)
- pip install weasyprint pymupdf
- python skills/anti-ai-poster-design/scripts/render_poster.py poster.html poster.png --width 1080 --height 1350
- # Audit the source: fonts, gradients, effects, containers, colors, text size, word count, clichés
- python skills/anti-ai-poster-design/scripts/poster_lint.py poster.html
- # Audit the render: calm space, edge busyness, palette size, hue families
- pip install pillow numpy
- python skills/anti-ai-poster-design/scripts/poster_pixels.py poster.png
- # Run the repo's own sanity tests
- python tests/run_tests.py
- ```
- The linter and pixel checker are heuristics. They catch objective tells; they cannot judge taste. A pass means "no known tell found", not "good design".
- ## Fonts
- Examples use free Google Fonts (Bebas Neue, Fraunces, Inter, Hind Siliguri, Anek Bangla). The templates load them from Google Fonts when online and fall back to installed or generic fonts. Install the fonts locally for offline rendering.
- ## How the skill was derived (and its limits)
- - The "AI" set was 12 photographed banners from a street in West Bengal, labelled AI-generated by the person who collected them; the reference set was 8 modern social-media posters. The labels are unverified.
- - Photographs add glare and noise, so measured busyness is overstated. Some tells (borders, pills, contact strips) are also typical of cheap banner printing in general, with or without AI.
- - The reference set is itself template-flavored, so the skill borrows *principles* (hierarchy, restraint, whitespace), not a look.
- - The original photos are **not** included: they show real businesses and phone numbers.
- Details and measurements: [`references/ai-tells-catalog.md`](skills/anti-ai-poster-design/references/ai-tells-catalog.md).
- ## Contributing
- Issues and pull requests are welcome, especially:
- - More tells (with a description and an example described in words, not a photo of a real business).
- - Palettes and font pairings for other scripts and regions (Devanagari, Tamil, Arabic, CJK...).
- - New layout skeletons and worked examples.
- - Better heuristics in `poster_lint.py` / `poster_pixels.py`, with a test in `tests/`.
- ## License
- MIT. See [`LICENSE`](LICENSE).
Ratings and comments
Sign in to rate this skill.
Sign inFree to use. If it’s helped, pass it on.