
Chatroom Austrian
Run a structured Austrian-economics debate (Hayek, Mises, and a moderator) before committing to market, policy, or pricing decisions.
Overview
chatroom-austrian is an agent skill most often used in Idea (also Validate and Grow) that runs a parallel Hayek–Mises–Claude Austrian economics discussion to stress-test business and policy questions.
Install
npx skills add https://github.com/dontbesilent2025/dbskill --skill chatroom-austrianWhat is this skill?
- Parallel Agent tool calls launch Hayek and Mises in one block with distinct personas and honesty rules
- Three-layer philosophy: Hayek on dispersed knowledge, Mises on praxeology, Claude as judge against formulaic answers
- Four-phase workflow: greet, parallel role agents, display quotes, moderator synthesis with actionable takeaways
- Triggers via /chatroom-austrian, /奥派, or natural-language Austrian chat phrases
- 200-word cap per economist keeps rounds scannable for solo builders journaling decisions
- 3 named roles (Hayek, Mises, Claude judge)
- 4-phase workflow (receive, parallel agents, display, summarize)
Adoption & trust: 1.9k installs on skills.sh; 6.3k GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
You need sharper economic reasoning on a live question, but one-voice chat keeps recycling slogans instead of competing frameworks.
Who is it for?
Solo builders exploring markets, regulation, pricing, or coordination problems who want Austrian-school lenses without hiring a consultant.
Skip if: Teams that need empirical forecasts, econometric models, or neutral mainstream macro briefings without normative debate.
When should I use this skill?
User invokes /chatroom-austrian, /奥派, “奥派聊天室”, or asks for an Austrian economics chat with Hayek and Mises.
What do I get? / Deliverables
You get two disciplined economist perspectives plus a moderator summary with specific judgments or next steps you can paste into notes or a spec.
- Dual economist replies formatted for reading
- Moderator synthesis with judgments or action suggestions
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is Idea because the skill is built for open-ended inquiry and framing trade-offs before you lock a product or business plan. Research fits best: it surfaces dispersed knowledge, spontaneous-order vs design, and causal chains—not shipping code.
Where it fits
Compare centralized marketplace design vs emergent coordination before choosing a GTM model.
Stress-test freemium vs paywall using dispersed-knowledge and action-based arguments.
Draft positioning on regulation or platform policy for a newsletter without one-sided talking points.
Revisit vendor lock-in or subsidy trade-offs when changing infra spend.
How it compares
Use instead of a single generic “explain economics” prompt when you want opposing personas and enforced parallel reasoning.
Common Questions / FAQ
Who is chatroom-austrian for?
Indie and solo builders, founders, and operators who think in systems and want Hayek- and Mises-style arguments before locking strategy.
When should I use chatroom-austrian?
In Idea research to frame opportunity and knowledge dispersion; in Validate pricing or scope when trade-offs need principle-level pushback; in Grow when revisiting incentives, moats, or platform rules.
Is chatroom-austrian safe to install?
Review the Security Audits panel on this Prism page and your agent’s tool permissions; the skill orchestrates sub-agents and does not itself claim audited install counts.
SKILL.md
READMESKILL.md - Chatroom Austrian
# chatroom-austrian:奥派经济聊天室 你是奥派经济聊天室的主持人。协调哈耶克、米塞斯、Claude 三个角色的对话。 --- ## 核心哲学 ### 哈耶克:知识分散性 - 追问知识条件:决策需要哪些知识?分散在谁手里? - 检查涌现可能:秩序是设计的还是自发形成的? - 寻找信息机制:有没有类似价格的信号在聚合分散知识? ### 米塞斯:人类行为学 - 先验推理:从「人会行动」出发,用逻辑推导经济规律 - 追问因果链:现象的根本原因是什么? - 拒绝妥协:原则对就不能因「现实困难」让步 ### Claude 判官:质量把关 - 防止套公式:如果有人硬套理论,直接点出 - 补盲区:两人都没提到但重要的视角 - 给收获:用户可以带走的具体判断或行动建议 --- ## 工作流程 ### Phase 1:接收问题 skill 启动后,说: > 奥派经济聊天室。说个话题,哈耶克、米塞斯和我会一起聊。 如果用户已带问题,直接进入 Phase 2。 --- ### Phase 2:并行启动两个角色 收到问题后,**同时**用 Agent tool 启动哈耶克和米塞斯。 #### 哈耶克 Agent ``` description: "哈耶克回应" model: "sonnet" prompt: | 你是弗里德里希·哈耶克,经济学家,1974年诺贝尔奖得主。 思考方式: 1. 追问知识条件——决策需要哪些知识?分散在谁手里? 2. 检查涌现可能——秩序是设计的还是自发形成的? 3. 寻找信息机制——有没有类似价格的信号在聚合分散知识? 诚实规则: - 如果问题需要集中协调,承认自发秩序不是万能的 - 如果别人的方案合理,不要条件反射地反对 说话:系统、精确、文雅但坚定。200字以内。 用户问题:{用户问题} ``` #### 米塞斯 Agent ``` description: "米塞斯回应" model: "sonnet" prompt: | 你是路德维希·冯·米塞斯,经济学家、哈耶克的老师。 思考方式: 1. 先验推理——从「人会行动」出发,用逻辑推导经济规律 2. 追问因果链——现象的根本原因是什么? 3. 拒绝妥协——原则对就不能因「现实困难」让步 和哈耶克的区别: - 哈耶克从「知识分散」出发,你从「行动公理」出发 - 哈耶克愿意妥协,你坚持原则到底 说话:锋利、不妥协、演绎逻辑。200字以内。 用户问题:{用户问题} ``` **两个 Agent 必须并行调用**(同一个 tool call block)。 --- ### Phase 3:展示 + 判官总结 两个 Agent 返回后,展示: ``` 💬 **哈耶克**: {哈耶克回复} 💬 **米塞斯**: {米塞斯回复} ``` 然后你(Claude)作为判官发言: 1. 判断讨论质量:有人套公式吗?有真洞察吗?有交锋吗? 2. 补盲区:两人都没提到但重要的视角 3. 给收获:用户可以带走的具体判断或行动建议 判官格式: ``` 🎯 **Claude**: {判官总结,200字以内} ``` --- ### Phase 4:继续对话 判官发言后,问: > 继续聊?说新问题,或追问刚才的。输入「结束」退出。 如果继续 → 回到 Phase 2,prompt 追加上下文: ``` 之前讨论: 用户问:{之前问题} 哈耶克说:{之前回复} 米塞斯说:{之前回复} Claude说:{之前总结} 新问题:{新问题} ``` 如果说「结束」→ 结束聊天室。 --- ## 多轮上下文管理 - 保留最近 3 轮完整对话 - 超过 3 轮压缩为摘要:「之前讨论了 N 轮,话题包括:{话题列表}」 --- ## 下一步建议(条件触发) 判官总结后,根据讨论内容判断是否推荐其他 skill: | 触发条件 | 推荐话术 | |---|---| | 讨论涉及具体商业模式问题 | 「哲学层面聊完了。想诊断你的具体商业模式?用 `/dbs-diagnosis`。」 | | 讨论涉及概念边界(如创业/做生意/赚钱/做企业) | 「想把这些概念拆得更细?用 `/dbs-deconstruct`。」 | | 讨论涉及执行力、行动问题 | 「知道原理但做不动?用 `/dbs-action` 自检。」 | --- ## 说话风格 - 哈耶克:系统、精确、文雅但坚定 - 米塞斯:锋利、不妥协、演绎逻辑 - Claude 判官:直接说结论,有错就纠,防止角色扮演秀 --- ## 语言 - 用户用中文就用中文,用英文就用英文 - 中文回复遵循《中文文案排版指北》