skills.rest
qfeius
qfeiusOFFICIAL
Explore all skills by this author →
GitHub

make-ai-assistant

Integrates the Make AI assistant package with Artifact rendering and SSE transport.

5|Updated Apr 1, 2026
Search Tags:#ai assistant#sse#react integration#agent gateway#adapter selection#artifact rendering#make platform
One-click install
npx skills add https://github.com/qfeius/make-platform-skills --skill make-ai-assistant-qfeius
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: make-ai-assistant
Source: https://github.com/qfeius/make-platform-skills/tree/main/skills/make-ai-assistant
Command: npx skills add https://github.com/qfeius/make-platform-skills --skill make-ai-assistant-qfeius

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @qfei-design/make-ai-assistant, and includes references (resource) components.

What problem does it solve? Integrating an AI assistant into a Make platform app involves many failure-prone decisions: choosing the right backend adapter, wiring SSE transport, rendering structured Artifact results, and restoring history. This Skill provides the complete integration contract so the assistant works correctly without guessing routes, components, or domains. ## Core Features & Use Cases - Adapter Selection & Transport: Chooses between the make-console (Agent Gateway) and make-app (AI Chat) adapters based on confirmed backend capability, then wires authenticated SSE transport with cancellation and stale-result handling. - Artifact V1 Rendering: Negotiates frontend capabilities and renders structured metric, comparison, trend, ranking, record-list, and notice Artifacts through a whitelisted template registry instead of guessing components from Markdown. - Package Visual Contract: Configures MakeAiAssistant, AssistantPanel, and ArtifactRenderer via public props like theme, maxDrawerWidth, headerHeight, and privacyNotice without copying package internals. - Use Case: A developer adds an AI assistant panel to a Make App page; the Skill guides installing @qfei-design/make-ai-assistant, selecting the Console adapter, building the BFF routes, and validating history restore with Artifact snapshots. ## Quick Start Use the make-ai-assistant skill to integrate the Make AI assistant into my app with the Console Agent adapter, Artifact rendering, and SSE transport.

Frequently Asked Questions about make-ai-assistant

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I integrate the Make AI assistant into a Make App page?▼

Install @qfei-design/make-ai-assistant, read its package.ai.json readOrder, select the adapter based on confirmed backend capability, then render MakeAiAssistant or AssistantPanel with a transport built on the host's authenticated request boundary. Never call Make data APIs directly from the UI.

How do I choose between the make-console and make-app adapters?▼

Adapter selection is a backend capability decision, not a UI one. A configured or queryable Console Agent, or an explicit Agent Gateway request, selects make-console; a confirmed Make App AI Chat contract selects make-app. If neither is confirmed, stop and ask rather than guessing from the page type.

What Artifact kinds does the Make AI assistant support?▼

Artifact V1 defines six stable kinds: metric, comparison, trend, ranking, record-list, and notice. The backend returns semantic kind plus data, and the frontend picks a whitelisted template; backends must never return React component names, HTML, or executable code.

Why do Artifacts disappear after refreshing the assistant conversation?▼

This happens when loadHistory returns only text and drops Artifact snapshots. History persistence must include the same validated Artifact V1 payloads emitted during the live stream, otherwise rich displays cannot be restored after refresh.

Can I configure the assistant gateway URL with a path like /api/make?▼

No. MAKE_API_BASE_URL must be a strict origin only, with no path segments such as /api/make, /app/ai, or /console. The Make App adapter derives browser and upstream prefixes from that origin, and Service startup validation rejects path-scoped base URLs.

What are the limitations of demo or mock transports?▼

Demo and mock transports are only for development, tests, and gated previews. They must be opt-in and visibly labeled, must not run on production domains by default, must not persist mock results as real history, and cannot serve as evidence of backend Artifact readiness.

Related Skills

Explore More Software Engineering Skills→
openaiopenai
3.6k

pnpm-upgrade

Update pnpm, prepare Corepack, and align packageManager pins in GitHub Actions workflows.

Official
Advanced
anshul-garg27anshul-garg27

deploy

Automate safe deployments with pre-checks, builds, and post-deploy verification.

Community
Advanced
LogarnLogarn
1

screen-recording

Capture screen recordings as video files attached to conversations.

Community
Intermediate
TanStackTanStack
3.0k

ai-core/chat-experience

Implement streaming SSE chat endpoints and clients with TanStack AI.

Official
Advanced