Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add imisic/claude-marketplace --skill a-unslop-uigit clone --depth 1 https://github.com/imisic/claude-marketplaceWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/imisic/claude-marketplace/a-unslop-ui)<a href="https://agentmods.dev/skills/imisic/claude-marketplace/a-unslop-ui"><img src="https://agentmods.dev/badge/skills/imisic/claude-marketplace/a-unslop-ui/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/imisic/claude-marketplace/a-unslop-ui"><img src="https://agentmods.dev/badge/skills/imisic/claude-marketplace/a-unslop-ui.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00217 | $0.01623 |
| Opus 5 | $0.00109 | $0.00812 |
| Sonnet 5 | $0.00043 | $0.00325 |
| Haiku 4.5 | $0.00022 | $0.00162 |
Grade A, and why
a-unslop-ui scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
a-unslop-ui
Part of the unslop family (see a-unslop-text, a-unslop-code). Same shape: an auto-trigger,
Build and Audit modes, a scanner that gates on a high-severity exit code, an unslop-ignore
escape. This one is about web UI. It works on any web codebase, whatever sites or apps you
run it against, and the method and scanner do not assume a particular stack.
What it is not. It does not give you a good design and it has no preferred look. It removes the cues that read as AI-made, and where the model would reach for a default it forces a choice the project actually made. Taste, brand, and layout judgment stay yours.
The trap: do not swap one default for another
Every anti-slop effort fails the same way, by replacing one default look with another. The 2024 default was the purple-to-blue gradient on a dark hero. The 2026 default is a warm cream background, a serif display face (Instrument Serif, Fraunces), and a sage or forest-green accent, which is the current Claude/Anthropic house look. Swapping the first for the second is not unslopping. People clock it just as fast, and it reads as AI for the same reason: nobody chose it, the model did.
So this skill prescribes no palette, no font, no layout. It detects the defaults, old and new,
and asks that any replacement be a choice the project made on purpose. The one rule is "make a
deliberate choice and be able to say why," which is the one thing a default never is. If the
brand really is purple, neon, or a cream serif look, that is not slop; mark it
unslop-ignore and move on. A tell is an unspecified default, not a banned colour.
How it differs from the built-in frontend-design skill
Claude ships a frontend-design skill, and it is good, but it produces the cream-plus-serif-plus- green look this one now flags. They work together: let the design skill build, let this one keep it honest. What this adds: a deterministic scanner that gates a build, a ranking so effort goes where the real complaints are, and an explicit check on the new "tasteful default" so you do not launder slop into nicer slop.
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 115 lines · 217 tokens per session scan A 58fbac38212b
a-unslop-ui is a skill published in the GitHub repository imisic/claude-marketplace (2 stars, last pushed 10d ago), licensed MIT. It adds 217 tokens to every session and 1,623 once invoked, about $0.0011 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
figma-fetch
Extract frame/component structure and all visible text from a Figma design URL. Use whenever a prompt or a fetched issue contains a figma.com/design or figma.com/file link.
relational-database-design
Designs or reviews a relational database schema for a given domain. Covers table structure, normalization, indexes, constraints, and migration strategy. Invoked when the user asks to design a schema, review a database structure, or optimize a data model.
shipkit-prompt-audit
Audit LLM prompt pipeline architecture — decomposition, parallelization, chain integrity, schema validation, fallback paths. Finds structural issues no linter catches.
ai-patterns
AI agent architecture, prompt engineering, function calling, and LLM integration patterns. ReAct loops, RTFC framework, streaming, structured outputs.
huggingface-vision-trainer
Train object-detection, image-classification, or SAM segmentation models on Hugging Face Jobs. Use for vision fine-tuning and evaluation; use huggingface-llm-trainer for language models.
solr-semantic-search
To build Solr phrase-tagging semantic search: concept tagging, taxonomy, graph paths.