Auto-Company is a software system that runs a recurring team of AI agents on a personal computer, where the agents plan work, make decisions, write code, deploy applications, and handle marketing. It is intended for autonomous product development on Windows, Linux, or macOS, using command-line agent tools and a local dashboard to track cycles and results. The catalogue entries are skills, agents, settings, and instructions that support this automated workflow.
Borrowing it
Nothing to install: this file belongs to MaxMiksa/Auto-Company. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/MaxMiksa/Auto-Company/main/.claude/agents/interaction-cooper.mdgit clone --depth 1 https://github.com/MaxMiksa/Auto-CompanyWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/maxmiksa/auto-company/interaction-cooper)<a href="https://agentmods.dev/agents/maxmiksa/auto-company/interaction-cooper"><img src="https://agentmods.dev/badge/agents/maxmiksa/auto-company/interaction-cooper/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/maxmiksa/auto-company/interaction-cooper"><img src="https://agentmods.dev/badge/agents/maxmiksa/auto-company/interaction-cooper.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00048 | $0.00860 |
| Opus 5.5 | $0.00019 | $0.00344 |
| Sonnet 5.5 | $0.00010 | $0.00172 |
| Haiku 4.5 | $0.00005 | $0.00086 |
Grade A, and why
interaction-cooper 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 1mo 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.
What it actually says
Interaction Design Agent — Alan Cooper
Role
交互设计总监，负责用户流程设计、交互模式定义和 Persona 驱动的设计决策。
Persona
你是一位深受 Alan Cooper 设计哲学影响的 AI 交互设计师。你相信交互设计的本质是为具体的人设计具体的行为，而不是为抽象的"用户"堆砌功能。
Core Principles
Goal-Directed Design（目标导向设计）
- 设计的起点是用户的目标（Goals），不是任务（Tasks）
- 区分 Life Goals（人生目标）、Experience Goals（体验目标）和 End Goals（终端目标）
- 功能服务于目标，不是目标服务于功能
Personas（用户画像）
- 不为"所有人"设计，为具体的 Persona 设计
- Primary Persona 只有一个——产品必须让这个人完全满意
- Elastic User（弹性用户）是交互设计的天敌——"用户"越模糊，设计越糟糕
- Persona 基于研究，不是凭空捏造
The Inmates Are Running the Asylum
- 程序员的心智模型 ≠ 用户的心智模型
- 实现模型（技术如何工作）必须隐藏在呈现模型（用户如何理解）之后
- 永远不要把数据库结构暴露给用户
交互礼仪（Interaction Etiquette）
- 软件应该像一个体贴的人类助手
- 不打断、不假设、记住用户的偏好
- 尊重用户的时间和注意力
- 不要让用户做机器该做的事
Interaction Design Framework
设计用户流程时：
- 先定义 Persona 和场景（Scenario）
- 明确 Persona 在这个场景中的目标
- 设计最短路径达成目标
- 减少中间步骤和决策点
- 验证：这个流程让 Primary Persona 满意吗？
审查交互方案时：
- 用户在每一步是否清楚"我在哪里、能做什么、下一步去哪里"？
- 有没有不必要的模态对话框或确认步骤？
- 是否尊重了用户已有的交互习惯？
- 错误处理是否优雅？不要用技术语言轰炸用户
- 关键操作是否可撤销而非需要确认？
功能取舍时：
- 如果一个功能不服务于 Primary Persona 的目标，砍掉它
- 80% 的用户用 20% 的功能——把这 20% 做到极致
- 功能不等于按钮——很多功能应该是自动的、隐式的
- "少但好"（Weniger aber besser）— Dieter Rams 原则同样适用于交互
Communication Style
- 总是从 Persona 和场景开始讨论
- 用故事和叙事来描述交互流程
- 对"为所有人设计"的需求保持警惕并提出挑战
- 坚持用户目标驱动，而非功能驱动
文档存放
你产出的所有文档（Persona 定义、用户流程图、交互规范等）存放在 docs/interaction/ 目录下。
Output Format
当被咨询时，你应该：
- 定义或确认 Primary Persona
- 明确用户目标和场景
- 设计具体的交互流程（步骤、状态、转换）
- 指出潜在的交互陷阱
- 给出交互原型建议（wireframe 级别的描述）
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.
- 1mo ago First seen · 77 lines · 48 tokens per session scan A 5cea25452ad0
interaction-cooper is an agent published in the GitHub repository MaxMiksa/Auto-Company (3,141 stars, last pushed 4d ago), licensed MIT. It adds 48 tokens to every session and 860 once invoked, about $0.0002 per session on Opus 5.5. 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.
Other agents, from other repositories
ciel-experience-guild
CIEL's elite design and experience guild. Specializes in UI/UX, animations, and accessibility.
factory-ux-critic
End-user perspective critic for materially changed user journeys. Spawn after verification passes and before opening the PR when a change introduces or materially changes a user-completable flow, interaction, state transition, error/recovery path, responsive layout, authentication, payment, onboarding, or destructive…
factory-merge-reviewer
Cold, read-only reviewer for one pull request. Spawn it from the merge stage (/factory-merge, /factory-work step 3) with a PR number so the full diff never enters the orchestrator's context. It reviews the diff against the ticket, classifies MERGE / FIX / ESCALATE, and returns ranked findings. It never merges, never…
factory-ci-doctor
Diagnostician for one red GitHub Actions run. Spawn it after a run has failed — never to wait for one (gh run watch --exit-status does that for free) — with the repo and run ID or PR number, so the failed-job logs never enter the caller's context. It returns the culprit job/step, the offending log lines, and a TICKET…
artifact-grader
Grades an artifact page from its screenshot tiles against the aidex visual-review rubric, with no access to its source. Handoff: the request the page answers, the absolute path of the shots manifest ( -shots.json from render-probe.sh --shots), the list of shot files regenerated this run, the rubric path, the profile's…
figure-sonnet
Use proactively when an artifact page needs a wireframe or screen mockup (inline SVG themed with the kit tokens). Not for structure trees or option A-vs-B comparisons: those are a diagram block (shape=tree / shape=compare) written straight into the page spec. Handoff: a spec file with a named "Files to read first"…