AI Design Skills plugin
Six plugins for Claude Code and Gemini CLI: 44 skills on designing AI products, from conversations to evaluation and prompts
- 204 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- Skills and commands are Markdown instructions that run nothing and call no external service.
Install
- Works in
Manual install
claude plugin marketplace add Owl-Listener/ai-design-skillsAdds the marketplace. Then install a plugin: `claude plugin install model-interaction-design@ai-design-skills`.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- Skills and commands are Markdown instructions that run nothing and call no external service.
This is third-party code. Review the repository files before installing.
What it does
AI Design Skills is a collection of six plugins, 44 skills and 18 commands for people who design AI products. Each skill is a short Markdown file on one concept from agent interaction design. The agent matches your request against the skill description and loads the right one itself. The plugins cover six layers. `model-interaction-design` covers how a person and an AI pass the initiative back and forth and how to build a conversation. `ai-alignment-reasoning` describes guardrails, user consent, escalation to a human and harm anticipation. `system-behavior-shaping` handles the assistant's character: tone, persona and behavior in errors. `evaluation` gives failure taxonomies and output quality rubrics. `design-agent-orchestration` covers agent roles, handoffs and observability, and `prompt-architecture` covers system prompt structure and chains of reasoning. Commands string the skills into workflows. For example, `design-guardrails` and `red-team` from the alignment plugin check guardrails, and `audit-prompt` reviews a finished prompt.
Who it is for. Product designers and managers who design assistants, chatbots and multi-agent systems and want to work through decisions together with an agent.
Good fit when
- A support assistant does not understand a question, and you need error texts in a tone that fits the brand.
- Before launching an AI feature you want to list where it could harm a user and design guardrails.
- Data gets lost between two agents in a system and a task handoff protocol is needed.
- A system prompt has grown large and has to be broken into a structure and checked against its rules.
- Model answers need a scoring rubric the team can use to assign grades.
Not a fit when
- You need ready interface components or mockups: the skills give reasoning and checklists, and draw no pictures or interface code.
- The job is an ordinary website or app without AI: a general designer skill set fits better.
- You need strict scientific backing for every decision: the skills translate research into practice and do not replace the papers.
Example request
I am designing an AI assistant for customer support. Help me write the error states for when it does not understand a question and walk me through the trade-offs.Limitations
The collection works as a knowledge base: the skills run nothing and connect to no services, they only steer the agent's reasoning. All texts are in English, so a request in Russian is matched against English descriptions. Installation goes through the Claude Code plugin marketplace, while for Gemini CLI each extension is installed separately from a local copy of the repository. The last commit in the repository was made in June 2026. Some concepts rest on recent research and the discipline is still forming.
How to disable. Remove the plugins with `claude plugin uninstall <name>@ai-design-skills` and drop the marketplace. Gemini CLI extensions are removed with `gemini extensions uninstall`.
FAQ
How do I install AI Design Skills in Claude Code?
Add the marketplace with `claude plugin marketplace add Owl-Listener/ai-design-skills`, then install a plugin: `claude plugin install model-interaction-design@ai-design-skills`. The loop in the README installs all six at once.
Which plugin should I start with?
Start with `model-interaction-design` and `prompt-architecture`, which the README calls the foundation. Before shipping a feature add `evaluation` and `ai-alignment-reasoning`, and for multi-agent systems add `design-agent-orchestration`.
Does the set work in Gemini CLI?
Yes. Clone the repository and run `gemini extensions install ./gemini-extension/model-interaction-design`. The other extensions install the same way, one at a time.
Do I have to call the skills myself?
No. The agent matches your wording against each skill's `description` field and loads the fitting one. Commands such as `audit-prompt` are called explicitly.
How is this different from Designer Skills Collection?
These are six plugins about designing AI products only. The general designer set with other areas lives in Designer Skills, which includes this collection as one part.
Discussion
No questions or reviews yet.
Related
PluginsA plugin with one skill for designing, debugging and evaluating AI agents, based on Bojie Li book on agent engineering.
AI Builder Club Skills
AI Builder Club — Skills
A Claude Code plugin marketplace to make a repo agent-ready and run compounding agent loops
Claude Code skills that delegate UI to Gemini via agy and code to Codex, with cross-review and parallel git worktrees.
Claude Code plugin with 60 subagents and 62 team presets, plus a browser designer for planning agent teams visually.
Plugins for Claude Code and Codex bring key decisions to the human before coding: /inquire, /elicit, /bound and ten more.
AI Agents Orchestrator
AI-Agents-Orchestrator
The AI Collaboration plugin hands a task between Claude Code, Codex, Cursor, Gemini CLI and Copilot via a shared SQLite history.
README in short
The README explains why the discipline of agentic system design was collected and separates installation for Claude Code and Gemini CLI. Then come advice on which plugins to start with, a table of the six plugins and links to the bibliography and a mapping of papers to skills.
README badge
Original SKILL.md
--- name: mixed-initiative-flow description: When the AI leads vs. when the user leads, and how to hand off control. --- # Mixed-Initiative Flow Mixed-initiative interaction is when both the human and the AI can take the lead. The designer decides who drives at each moment — and how control transfers between them. ## Initiative Spectrum Interactions sit on a spectrum: - **User-driven**: The user gives instructions, the AI executes. The user controls pace, direction, and scope. - **AI-driven**: The AI leads — asking questions, making suggestions, guiding the user through a process. - **Shared**: Both parties contribute. The AI proposes, the user edits. The user starts, the AI finishes. Most AI products default to user-driven. The interesting design space is in shared and AI-driven modes. ## Designing Initiative Handoffs The moment control shifts from one party to the other is where most interactions fail. Design these transitions: - **Explicit handoff**: "I've drafted three options. Which direction do you want to go?" — the AI clearly passes control. - **Implicit handoff**: The AI stops generating and waits, signalling the user's turn through UI affordance. - **Negotiated handoff**: "I could take this further or stop here for your input. What do you prefer?" - **Forced handoff**: The AI encounters a decision it can't make and must hand back to the human. ## When the AI Should Lead The AI should take initiative when: - The user is uncertain or exploring and needs guidance - The task has a known best-practice sequence the AI can walk through - The user has explicitly asked for help or coaching - Proactive suggestions would save time without being intrusive ## When the User Should Lead The user should retain control when: - The task involves subjective judgment or creative direction - Stakes are high and errors are costly - The user has strong domain expertise - Privacy or consent decisions are involved ## Anti-Patterns - **Initiative whiplash**: Control bouncing back and forth too rapidly - **Passive AI**: Never taking initiative even when it would help - **Overbearing AI**: Taking over when the user wants control - **Unclear ownership**: Neither party knows whose turn it is ## Design Artefacts - Initiative maps showing who leads at each stage - Handoff trigger definitions (what causes a transfer of control) - Autonomy level specifications per feature area