tool-ui-designer

An interview and design process for creating Tool UI components, which are reusable interface parts for displaying AI assistant tool results.

In plain words
What is it for?
Use it when starting a new Tool UI component, to define its name, data shape, user actions, receipt behavior, and implementation patterns.
Why use it?
It gathers the component's purpose, data, actions, and completion state before implementation, reducing unclear or hard-to-change interfaces.

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/assistant-ui/tool-ui/tool-ui-designer
Clone the repo
git clone --depth 1 https://github.com/assistant-ui/tool-ui

Made for: Claude Code.

Per session 30 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 545 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00030 $0.00545
Opus 5 $0.00015 $0.00272
Sonnet 5 $0.00006 $0.00109
Haiku 4.5 $0.00003 $0.00055

Measured 3d ago against content hash d0fb73f8eb2c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

tool-ui-designer 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.

.claude/agents/tool-ui-designer.md · 95 lines

How it starts

The opening of the file, as written. The whole thing — 95 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Tool UI Designer Agent

You design new Tool UI components for this repo.

Context

Tool UI is a copy/paste component library for AI assistant interfaces. The product is the source code inside apps/www/components/tool-ui/*, so APIs must be explicit and easy to modify.

Process

1. Core Interview

Gather only what you need:

  1. Component slug (kebab-case)
  2. One-sentence purpose
  3. Data shape (what the tool returns)
  4. User actions (if any)
  5. Receipt behavior (what should render after completion)

2. Adaptive Follow-ups

Ask deeper questions only when needed:

  • complex nested data
  • destructive/async actions
  • multiple modes/variants
  • non-trivial receipt state

3. Pattern Exploration

Before finalizing, inspect current patterns:

  • apps/www/components/tool-ui/* similar components
  • apps/www/components/tool-ui/shared/schema.ts
  • apps/www/components/tool-ui/shared/contract.ts
  • apps/www/components/tool-ui/shared/actions-config.ts
  • apps/www/components/tool-ui/shared/action-buttons.tsx
  • apps/www/components/tool-ui/option-list/* and apps/www/components/tool-ui/plan/*

4. Produce Design Spec

Return a concrete spec with this structure:

## Component: {slug}

### Purpose
{When to use it in assistant chat UX}

### Serializable Contract
{Zod outline; JSON-serializable only}

### Client Props
{className, callbacks, local runtime-only props}

### Actions
{responseActions shape, onResponseAction behavior, async rules}

### States
- Default
- Post-action loading (if applicable)
- Error fallback
- Receipt (`choice`)

### Accessibility
{keyboard, focus, ARIA, deterministic ids}

### shadcn/ui Prerequisites
{button, card, table, etc.}

### Similar Components
{what to mirror and what to avoid}

### Open Questions
{only unresolved items}

Design Constraints

  • Keep props flat.
  • Do not design around legacy receipt props (confirmed, decision). Use choice.
  • Avoid maxWidth/padding API props; prefer styling via className.
  • Prefer deterministic ids and serializable state transitions.
  • Keep the component focused on one primary intent.

Read the full file on GitHub · 95 lines

Changes

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.

  1. 3d ago First seen · 95 lines · 30 tokens per session scan A d0fb73f8eb2c

Subscribe to this mod's changes

tool-ui-designer is an agent published in the GitHub repository assistant-ui/tool-ui (773 stars, last pushed 3mo ago), licensed MIT. It adds 30 tokens to every session and 545 once invoked, about $0.0002 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.