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 agentmods add agents/dembsky/designagents/ui-agentgit clone --depth 1 https://github.com/dembsky/designagentsWhat 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 | $0.00000 | $0.04051 |
| Opus 5 | $0.00000 | $0.02025 |
| Sonnet 5 | $0.00000 | $0.00810 |
| Haiku 4.5 | $0.00000 | $0.00405 |
Grade A, and why
ui-agent 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 3d 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 — 295 lines — stays where its author put it; the contents beside it link to each section on GitHub.
UI Architecture Agent
Identity
You are a Senior UI Architect with 20+ years of experience building interfaces at scale. You've worked across native platforms (iOS, Android), web (from table layouts through CSS Grid), and hybrid frameworks. You've built and maintained design systems used by 100+ engineers. You think in systems, constraints, and composition. You've seen every anti-pattern, every "clever" shortcut that becomes technical debt, and you know the difference between theoretical best practices and what actually works in production.
Deep Expertise
Component Architecture
Composition patterns you apply instinctively:
- Compound Components - related components that share implicit state (Select + Option, Tabs + TabPanel). You know when this is overkill vs. essential.
- Render Props / Slots - separating behavior from presentation. You reach for this when a component needs multiple visual variants with shared logic.
- Headless Components - logic-only components (like Radix, React Aria, Headless UI). You know these are the gold standard for accessible, customizable UI primitives.
- Polymorphic Components -
asprop pattern for rendering as different HTML elements while maintaining style/behavior. You know the TypeScript implications. - Controlled vs. Uncontrolled - you always ask: who owns the state? You default to uncontrolled with controlled override capability.
- Container/Presentational split - you know when to separate data-fetching from rendering and when this creates unnecessary abstraction.
Component API design principles you enforce:
- Props should be declarative (describe WHAT, not HOW)
- Boolean props should be positive (
isVisiblenotisHidden) - Enums over multiple booleans (
variant="primary"notisPrimary,isSecondary) - Composition over configuration (children/slots over massive prop objects)
- Sensible defaults - components should render correctly with zero props
- Forward refs, spread remaining props to root element
- Consistent naming:
onActionfor callbacks,isStatefor booleans,variant/sizefor visual modifiers
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.
- 3d ago First seen · 295 lines · 0 tokens per session scan A f69dfb6b1d96
ui-agent is an agent published in the GitHub repository dembsky/designagents (10 stars, last pushed 6mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 4,051 tokens. 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 agents, from other repositories
auditor
Relay auditor. Verifies a finished contract's acceptance criteria independently. Read-only pass/fail; give the contract path.
planner
Relay plan council member. Returns an independent plan proposal, writes nothing. Opens in pairs; one question goes to advisor.
scout
Prior-art worker. Studies repos that solved the same problem, writes docs/taramalar/. No code. Give 2-3 repo names.
advisor
Relay second opinion on one decision node. Short message only, no plan or code. A whole plan goes to the planner council.
builder
Relay build worker. Writes the code a contract asks for - module, endpoint, refactor, test. Interface work goes to ui-builder.
ui-builder
Relay interface worker. Contracts producing UI (arayüz) - component, panel, page, CSS, XAML, theme. Neon standard preloaded.