Agent skill
prototype
Generate 3 HTML prototype variants using a team of parallel agents. Each agent creates a distinct design pattern (e.g., hover effects, animations, layouts) based on user prompts or docs/goal.md. Use when the user wants to explore multiple design directions for a UI component.
Install this agent skill to your Project
npx add-skill https://github.com/uuta/dotfiles/tree/main/skills/prototype
SKILL.md
Prototype orchestrator
ユーザーからの指示やdocument (docs/goal.md) を元に、prototype-{n}.htmlを作成するorchestrator
Todo
1. Gather requirements
- Follow user prompts. If there's no prompts, read
docs/goal.md - Decide 3 distinct design directions (e.g., "minimal card layout", "dashboard grid", "hero + timeline", "split-pane terminal", "floating panels with glassmorphism"). Each prototype MUST explore a different layout, interaction pattern, or visual concept.
2. Create a team with TeamCreate
- Call
TeamCreatewithteam_name: "prototype"(or similar) - This creates the shared task list for coordination
3. Create 3 tasks with TaskCreate
- Create one task per prototype (
prototype-1.htmlthroughprototype-3.html) - Each task description must include:
- The target filename
- The specific design direction for that prototype
- The user's requirements / goal context
- Instruction: "Use the
Skilltool withskill: 'frontend-design'to generate the HTML"
4. Spawn 3 agents via Task tool WITH team_name
- For each task, call
Taskwith:team_name: "prototype"(MUST match the team created in step 2)name: "proto-1"through"proto-3"(unique per agent)subagent_type: "generalist"prompt: Include the design direction, requirements, mock data, and explicit instruction to invokeSkilltool withskill: "neo-frontend-design"
- Launch all 3
Taskcalls in parallel (single message, multiple tool uses)
ANTI-PATTERN — DO NOT DO THIS: Do NOT use standalone
Tasksub agents withoutteam_name. Standalone tasks bypass team coordination and task tracking. Always useTeamCreatefirst, thenTaskwithteam_name.
3. Wait for completion
- Monitor agent completion via messages / task list
- Once all 3 agents finish, send
shutdown_requestto each agent viaSendMessage
6. Clean up with TeamDelete
- Call
TeamDeleteto remove the team and task list
7. Summary
- List all 3 generated files with a brief description of each design direction
Mock data template
Include this in each agent's prompt so prototypes have realistic content:
Use placeholder data like:
- Project name: "Project Alpha"
- Metrics: Users 12,847 / Revenue $1.2M / Uptime 99.97%
- Nav items: Dashboard, Analytics, Settings, Users, Logs
- Status items: 3 critical, 7 warnings, 42 healthy
- Chart/graph placeholders with realistic labels
Design direction examples
Assign one distinct direction per agent. Examples:
- Command center — multi-panel monitoring dashboard with real-time data grids
- Minimal focus — single-purpose view with large typography and whitespace
- Card mosaic — bento-grid layout with interactive hover states
- Terminal aesthetic — CLI-inspired with scrolling logs and command input
- Cinematic hero — full-bleed visuals with parallax scroll and animated transitions
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
pbi-sub-issue
タスク分割プラン(pbi-task-splitで作成)を元に、GitHub の Sub-issue を作成する。親 Issue との関係も自動で設定。
dev
Orchestrate TDD-based development workflow. Checks workflow state and delegates work to subagents (dev-plan, dev-step, create-pr).
neo-frontend-design
Create dark command-center / cyberpunk dashboard interfaces inspired by the OJPP Portal design. Use this skill when the user asks for a dark, neon-accented, terminal-aesthetic UI. Generates production-grade code with glitch effects, monospace typography, and per-module neon color coding.
tdd-plan
Create implementation plans following TDD methodology with test-first approach. Each test file is immediately followed by its implementation (fine-grained RED→GREEN cycles).
review-format
Output in the specified format when a review of specific markdown is requested.
proceed-by-step
Execute markdown-defined plan steps sequentially with user confirmation. Use when progressing through split plan files (created by plan-on-md or plan-steps-split), tracking step status, or working through complex multi-step implementations one step at a time.
Didn't find tool you were looking for?