Agent Skills
› langchain-ai/design-agent
› design-agent-onboarding
design-agent-onboarding
GitHub指导用户安装配置 Design Agent,涵盖环境验证、模型与媒体密钥设置、品牌及 Paper 集成、沙盒测试及部署选型。全程不暴露凭证,通过诊断工具确保环境就绪,为后续定制和评估做准备。
触发场景
需要安装或初始化 Design Agent
配置 AI 模型或第三方 API 密钥
连接品牌文件或 Paper 服务
排查 Agent 启动或运行错误
安装
npx skills add langchain-ai/design-agent --skill design-agent-onboarding -g -y
SKILL.md
Frontmatter
{
"name": "design-agent-onboarding",
"description": "Install and configure Design Agent, add a brand, connect Paper, choose MDA or self-hosting, and diagnose setup without exposing credentials."
}
Set up Design Agent
Work through these with the person, in order. Never print a key or token; report missing settings by name. Do not deploy, connect accounts or publish anything until the person asks.
- Install and prove the computer. Optional before a first MDA deploy, which needs no Docker.
uv sync, thenuv run design-agent sandbox build(Docker; a cold build downloads about 2 GB and takes several minutes) anduv run design-agent demo: a scripted run renders the sample brand's launch card in three sizes and a 4-second teaser with no keys. If it fails,uv run design-agent doctorsays what is missing. - Model.
uv run design-agent config set DESIGN_AGENT_MODEL=provider:model, then the person runsuv run design-agent config set ANTHROPIC_API_KEY(orLANGSMITH_GATEWAY_API_KEYfor alangsmith:model) in their own terminal and types the key hidden; never take a key in chat. Defaults and other roles:docs/models.md. Re-rundoctor, then, with their yes,uv run design-agent test model:doctoronly checks that a key is set, the test makes one cheap, billed call with it. - Media keys (optional).
OPENAI_API_KEY(image generation),ELEVENLABS_API_KEY(voice, music, sound effects),GROQ_API_KEY(fast transcription). Each tool says plainly when its key is missing. - Brand. Follow the
brand-onboardingskill to buildworkspace/brand/from what the person has (a brand guide, a website, files, or a Paper file). Until then the sample brand is used. - Paper (optional).
PAPER_MODE=localwhen Paper Desktop runs on this machine;bridgedfor a deployed agent (seedocs/partners/paper.md).offis fine: boards render in the sandbox. - Try it.
uv run design-agent ask "Make a LinkedIn card announcing …". Finals land inworkspace/out/<thread>/, the last line estimates what the run cost, and--thread <id>continues a conversation. - Choose where it runs. Managed Deep Agents (managed runtime, sandboxes, Slack app and schedules; one command) or self-hosted (Docker Compose with Postgres and their own Slack app). The design-agent-deploy skill takes it from there, including verifying with a real request.
- What next. Teach it the deliverable the team asks for every week (design-agent-customize), and measure changes with design-agent-evals.
版本历史
- 8134d7c 当前 2026-11-20 21:40


