Taste Skill
Leonxlnx/taste-skillAn anti-generic-design skill that gives the agent an art direction, open-licensed photos and a headless-browser check loop for web pages and app screens.
Recommended
Check first Depth 4/5
Monocolor Editorial Print (mono-color) Claude skill
A skill that turns a theme or photo into a one-ink or two-ink editorial print image: halftone, warm paper, strong type and generous negative space.
Our verdict Recommended
A narrow but unusually well-specified style skill with scripts, templates and tests; it only helps if you want this particular look.
What Monocolor Editorial Print (mono-color) does
mono-color produces a single visual style: an original poster, zine page or packaging image that looks printed with one or two inks on a neutral paper stock. Given a theme, a sentence, an object or a supplied photo, the agent extracts the subject, intent and any wording, then recombines a stored visual system into a new composition rather than copying a reference. The default output is controlled two-ink with one dominant plate and a narrow accent; an explicit one-ink or monochrome request stays pure one-ink. Contemporary editorial is the default and a vintage look only appears when requested.
Each run is meant to deliver three things: a final image-generation prompt, a generated raster image, and a short recipe note naming palette, layout family, type pairing and print process. Files are saved under a folder on the Desktop when a path is available. The skill file is about 34 KB and is backed by five reference documents, a design-system catalog of tokens, typography and composition rules, four Python scripts, 29 template files, a few tests and a validation workflow. It needs an image-generation capability in the host agent to produce the image; without one it stops at the prompt.
Good for
- Designers who want a consistent risograph or halftone editorial look
- Posters, zines and packaging mockups from a short brief
- Teams that need a documented, reusable visual grammar
Not for
- Full-color photography or brand-matching work
- Hosts with no image-generation tool
- Reusing the bundled example images commercially without permission
What's inside depth 4/5
A 34 KB skill with 5 references, 4 scripts, 29 templates, 3 tests and a validation workflow in CI.
- Skills
- 1
- Reference docs
- 5
- Scripts
- 4
- Tests / evals
- 5
- Templates
- 29
- Avg SKILL.md
- 33.5 KB
Counted from the repository file tree. How we score depth.
Dependencies and costs
- An image-generation tool or model available to the host agent
- Python for the helper scripts
Security review
Review before installingA visual-style skill that turns a theme into a one- or two-ink editorial image prompt and image, backed by JSON design-system catalogs and four small Python scripts that render example boards locally or validate the catalogs. It has no allowed-tools frontmatter and no network code; SKILL.md says to save generated files under ~/Desktop/Claude skills/mono-color/.
Worth knowing
- Relies on a CLI or package we could not reviewBoth poster/board build scripts need the rsvg-convert binary and call it with subprocess.run using a fixed argument list; the SKILL.md also expects an image-generation tool in the host agent.scripts/build_design_system_board.py:257
What we checked
The repository (68 files) on Oct 29, 2026. Not a full review: it relies on code outside the repository: rsvg-convert (librsvg) binary used by the build scripts; The image-generation capability of the host agent that the skill relies on to produce the final raster image.
NVIDIA SkillSpector static scan: 5 pattern hits (Output Handling 2, Dangerous Code Execution 2, MCP Least Privilege 1). The two HIGH 'unvalidated output injection' hits and the MEDIUM subprocess hit are the rsvg-convert calls in the two build scripts, which pass a fixed argument list and no user-controlled shell string. The MCP least-privilege note only says SKILL.md declares no allowed-tools while scripts use subprocess.
A review, not a certification. Skills change after we check them — read the files before installing.
License
MIT for code, instructions and scripts; the original example images are copyrighted by the author and not covered by the MIT license, and third-party research references remain with their owners.
Can it run unattended?
Not really. It is a one-shot creative generator with gate checks, not an unattended loop.
Alternatives to Monocolor Editorial Print (mono-color)
Hallmark
Nutlope/hallmarkA design skill from Together AI that deals a theme from 50 options, builds pages around the visitor's goal and lints the result in a real browser for AI-looking tells.
Recommended
Check first Depth 5/5
FAQ
What is the mono-color skill?
It is an agent skill that generates one-ink or two-ink editorial print images with halftone textures, neutral paper and bold typography. It outputs a prompt, an image and a short recipe note.
How do I install mono-color?
Copy the repository into your agent's skills folder or install it with the skills CLI. Your agent also needs an image-generation capability to produce the final image.
Can I use the example images commercially?
The code and instructions are MIT, but the README says the original examples are the author's copyright and are not covered by that license. Treat them as samples, not assets.
Sources: GitHub repository (stars, license, file tree, last push Sep 2, 2026). Review based on the README and SKILL.md files as published.