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 metaspartan/cybara --skill adversarial-ux-testgit clone --depth 1 https://github.com/metaspartan/cybaraWrote 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/metaspartan/cybara/adversarial-ux-test)<a href="https://agentmods.dev/skills/metaspartan/cybara/adversarial-ux-test"><img src="https://agentmods.dev/badge/skills/metaspartan/cybara/adversarial-ux-test/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/metaspartan/cybara/adversarial-ux-test"><img src="https://agentmods.dev/badge/skills/metaspartan/cybara/adversarial-ux-test.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00028 | $0.00428 |
| Opus 5 | $0.00014 | $0.00214 |
| Sonnet 5 | $0.00006 | $0.00086 |
| Haiku 4.5 | $0.00003 | $0.00043 |
Grade A, and why
adversarial-ux-test 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 10d 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.
What it actually says
Adversarial UX Test
Use this before demos, launches, onboarding changes, or major UI releases when the user wants friction found by a realistic hard-to-please persona.
Workflow
- Define one persona:
- Role, age range, technical comfort, core task, frustration triggers, and voice.
- Browse the app as that persona:
- First impression.
- Core workflow.
- Error recovery.
- Readability and accessibility.
- Speed and perceived effort.
- Terminology.
- Navigation and return paths.
- Record evidence:
- Screenshots for each major issue.
- Console errors.
- Network failures.
- Click count for the core task.
- Write the persona rant:
- In character, with concrete quotes and page references.
- Apply the pragmatism filter:
- Red: real UX bug.
- Yellow: valid but lower priority.
- Green: feature request with strong value.
- White: persona noise.
- Produce tickets only for Red and Green items.
Rules
- One persona per session.
- Test the core workflow before touring settings.
- Do not create more than 10 tickets.
- Separate entertaining persona feedback from objective product findings.
- Do not turn "I hate computers" into product work unless it exposes a real accessibility or workflow issue.
Report Shape
Persona: <name and constraints>
Task: <core task>
Verdict: <would they keep using it?>
Persona feedback:
- "<quote>" - <what happened>
Filtered findings:
- RED: <issue, evidence, suggested fix>
- YELLOW: <issue>
- GREEN: <request>
- WHITE: <noise>
Tickets:
- <title> - <evidence> - <fix>
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.
- 10d ago First seen · 64 lines · 28 tokens per session scan A b9e22cef9990
adversarial-ux-test is a skill published in the GitHub repository metaspartan/cybara (28 stars, last pushed yesterday), licensed MIT. It adds 28 tokens to every session and 428 once invoked, about $0.0001 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 2026-08-30.
Other skills, from other repositories
uiux-design
Skill "uiux-design" from KornLabs/truss, covering design, when to use, sub-skill routing, logo design (built-in) and logo: generate design brief.
uiux-ui-ux-pro-max
Skill "uiux-ui-ux-pro-max" from KornLabs/truss, covering ui/ux pro max - design intelligence, when to apply, rule categories by priority, running the search tool and workflow.
composio-canvas-design
Skill "composio-canvas-design" from KornLabs/truss, covering design philosophy creation, the critical understanding, how to generate a visual philosophy, philosophy examples and essential principles.
superpowers-brainstorming
Skill "superpowers-brainstorming" from KornLabs/truss, covering brainstorming ideas into designs, anti-pattern: "this is too simple to need a design", checklist, process flow and the process.
uiux-ui-styling
Skill "uiux-ui-styling" from KornLabs/truss, covering ui styling skill, reference, when to use this skill, core stack and component layer: shadcn/ui.
anthropic-frontend-design
Skill "anthropic-frontend-design" from KornLabs/truss, covering frontend design, ground it in the subject, design principles, restraint and self-critique and more on writing in design.