Hubbestaiskill.com

Home / All Skills / Workflow & Core Tools / dbs-chatroom-austrian

Workflow & Core Tools · dontbesilent2025/dbskill

dbs-chatroom-austrian

A multi-role discussion from an Austrian Economics perspective featuring Friedrich Hayek, Ludwig von Mises, and Claude. Used when the user requests to enter the Austrian Economics chatroom or analyze a problem using Austrian economics.

L1 Basic · Source registration onlyV1 Commit-PinnedDual-pass Audit

Author dontbesilent2025GitHub dontbesilent2025/dbskill ↗Stars 10121License CC BY-NC 4.0（仓库根 LICENSE 首行「Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)」，条款为 Attribution + NonCommercial，Copyright (c) 2026 dontbesilent。**非商业授权**：个人/学习/研究可直接使用，商业用途需单独授权；GitHub API license.spdx_id 为 NOASSERTION（CC 系非 SPDX 标准标识））Commit 8b8e33f1ec

Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Fixed three-character lineup, each bound to a specific methodology with a hardcoded tone: Hayek represents the dispersion of knowledge, Mises represents praxeology, and Claude serves as the judge for quality assurance.

skills/dbs-chatroom-austrian/SKILL.md

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。

The core philosophical section converts their division of labor into an actionable checklist: three rules for Hayek (probing knowledge conditions / examining emergent possibilities / seeking information mechanisms), three rules for Mises (a priori deduction / tracing causal chains / refusing compromise), and three rules for Claude (preventing formulaic answers / addressing blind spots / providing takeaways). These serve both as thinking instructions and as criteria for the judge.

The definitions of the two character Agents (description / model / prompt) are fully embedded in SKILL.md. Their prompts include 3 thinking modes, honesty rules, and speaking styles, and require parallel invocation within the same tool call block.

skills/dbs-chatroom-austrian/SKILL.md

**两个 Agent 必须并行调用**（同一个 tool call block）。

Hayek's honesty rules include 'If a problem requires centralized coordination, admit that spontaneous order is not omnipotent'; Mises explicitly states differences from Hayek ('Hayek starts from "dispersed knowledge", you start from the "action axiom"'; 'Hayek is willing to compromise, you stick to principles to the end'), ensuring the two do not merely repeat the same talking points.

The model is hardcoded as sonnet in the character definitions (a host-side dependency, hardcoded in the prompt code block).

skills/dbs-chatroom-austrian/SKILL.md

model: "sonnet"

Both character Agents include `model: "sonnet"` and set the output limit to under 200 words. This is one of the few places where a host/model choice is written into the skill: if model names differ across hosts (Codex, Doubao, etc.), this parameter may become invalid or trigger errors. This is a host-adaptation issue and does not alter the skill's own behavioral classification.

The judge (Claude) has three responsibilities and restricts output length to 200 words: assess quality (check for formulaic answers/genuine insights/clashes), address blind spots, and provide takeaways.

skills/dbs-chatroom-austrian/SKILL.md

防止套公式：如果有人硬套理论，直接点出

The judge format is '🎯 **Claude**: {judge summary, under 200 words}', and the speaking style section mandates 'State conclusions directly, correct errors immediately, and prevent roleplay posturing'—the judge's purpose is to prevent the characters from devolving into pure theatrical performance.

Continuing conversations appends context using a fixed template (previous question + responses from both characters + judge summary + new question), fostering progressive discussion rather than repeated statements.

skills/dbs-chatroom-austrian/SKILL.md

> 继续聊？说新问题，或追问刚才的。输入「结束」退出。

The Phase 4 context appending template is 'Previous discussion: User asked: {previous question} Hayek said: {previous response} Mises said: {previous response} Claude said: {previous summary} New question: {new question}'; the exit keyword is '结束' (different from '结束聊天室' in dbs-chatroom).

Multi-turn memory features an explicit compression strategy: retain the last 3 turns of full dialogue, compressing anything beyond into a one-sentence topic summary.

skills/dbs-chatroom-austrian/SKILL.md

- 保留最近 3 轮完整对话

The associated compression format is 'Previously discussed N rounds, topics included: {topic list}'. This approach controls context expansion at the prompt layer, consistent with the skill's zero-disk-write design (persisting no conversation logs locally).

2Core Capabilities

01Zero-configuration launch: parallel chat starts immediately upon receiving a user topic, without character selection or recommendation confirmation steps

02Fixed parallel dual-perspective output: Hayek (dispersed knowledge/spontaneous order) and Mises (action axiom/deduction), each limited to under 200 words

03Fully usable embedded character Agent definitions (description + model + 3 thinking modes + honesty rules + speaking style)

04Claude judge summary: assesses quality (formulaic vs. insightful/clashing), addresses blind spots, and delivers actionable takeaways

05Follow-up and exit mechanisms: appends context using a fixed template to continue discussions; type '结束' to exit

06Multi-turn context management strategy: retains the last 3 full turns of dialogue, compressing older turns into topic summaries

3External Dependencies

Kind	Dependency / Endpoint
cli	宿主 Agent tool（并行调用两个角色子代理；由宿主提供，非第三方依赖）
package	Anthropic Claude「sonnet」模型（写死在角色 Agent 定义里，宿主侧依赖）

4Recommended Governance & Evidence

Suggested Controls for Safe Operation

For Agent Runtime

Can be auto-loaded

For Administrators

Register source and version

Evidence & Triggers (File Paths & Excerpts)

SKILL.md: Pure prompt instructions and rule text, without script execution or external side effects

Audit Notes: . Self-verified against Tier 1 of the Batch 4 final criteria: the dbs-chatroom-austrian directory contains only SKILL.md and agents/openai.yaml, with no scripts, references, templates, or data files. It contains no executable scripts, contacts no network endpoints, reads no credentials, and performs no disk writes. Its sole external runtime action is launching two character Agents in parallel via the host's Agent tool—a host-native capability rather than a third-party service—and prompts are not forwarded to an external generation backend. Host-side considerations (not counted against safety tiering): SKILL.md hardcodes `model: "sonnet"` in the character Agent definitions, which may degrade or error if the host lacks that model identifier (a host/model availability issue). Additionally, 'multi-turn context management (retaining the last 3 turns)' is an artificial prompt-level convention rather than local disk storage. The only remaining risks stem from the prompt layer and roleplay outputs (see risks).

5Second-Pass Verification

[ok] 无脚本/无网络/无凭证/无写入（green 档自证） — 目录仅 SKILL.md + agents/openai.yaml；token 扫描中 http|curl|wget|API_KEY|token|secret|\.env|bash|exec|python 零命中；命中的 Agent tool / model 属宿主能力；无写盘指令。
[ok] model: "sonnet" 是否构成外部服务依赖（档位核查） — 它写在宿主 Agent tool 的角色定义里，是对宿主/模型的选择，而非本 skill 自带脚本去调用某端点；按「判级对象＝skill 自身目录、宿主行为不判级」的口径不升档，已在 reason 与 external_deps 中如实登记为宿主侧依赖。
[ok] 「并行调用」要求逐字存在 — Phase 2 结尾为「**两个 Agent 必须并行调用**（同一个 tool call block）。」，两个角色定义各自给出 description/model/prompt，未见串行或省略。
[ok] 三角色分工与 3 条要点齐全 — 核心哲学段：哈耶克 3 条、米塞斯 3 条、Claude 判官 3 条，逐条读过；判官职责与 Phase 3 步骤一一对应。
[ok] 多轮记忆策略与退出词 — 「保留最近 3 轮完整对话」+ 压缩格式逐字存在；退出指令为「结束」→「结束聊天室。」（注意与 dbs-chatroom 的「结束聊天室」措辞不同，已在 how_it_works 注明）。
[ok] 与 marketplace 条目描述一致 — marketplace.json 描述「哈耶克 × 米塞斯 × Claude 三人对话。奥派经济学视角的多角色讨论。」与 SKILL.md 一致。
[ok] 元数据（license/stars/last_push/installs/commit） — GitHub API：stars 10121、pushed_at 2026-09-07T03:48:59Z、license.spdx_id NOASSERTION（根 LICENSE 为 CC BY-NC 4.0，非商业）；本地 HEAD = 8b8e33f1ecaed8cee606fe950c4426b525ead314 = 任务 pin。
[discrepancy] meta.installs 值域校正（批4 质检 2026-09-15） — 原记录为 skills.sh 仓库级聚合徽标（579.6K（skills.sh 徽标 /b/dontbesilent2025/…），并非本 skill 的单体安装量；站内惯例只记可定位的单 skill 数值 → 校正为 null（其他 25 份 dbs 报告本就为 null，修此 8 份以齐平）

6Verdict & Scenarios

Zero-configuration and low friction: Characters and methodologies are fixed; discussions begin as soon as a topic is provided, making it ideal for repeated similar analyses.

Institutionalized opposition: Distinct starting points and compromise tendencies are explicitly written into both prompts, ensuring genuine intellectual clashes rather than redundant restatements.

The judge's duties are strictly defined with anti-performance clauses, converging outputs into actionable takeaways.

Quantified control over context expansion (retains 3 turns + summary compression) prevents long sessions from accumulating context indefinitely.

Embedded character definitions are fully self-contained and reproducible, requiring only model parameter adjustments when migrating across hosts.

Best for: Ideal for users looking to stress-test a decision or proposition quickly from an Austrian Economics perspective: providing a topic generates comparative arguments from two complementary frameworks (dispersed knowledge vs. the action axiom) alongside a judge's summary. It is especially suited for qualitative discussions on market mechanisms, price signals, spontaneous order, and interventionist policies. It also serves as a minimal exemplar of a 'fixed-character debate skill' (featuring all four elements: embedded role definitions, parallel calls, judge convergence, and 3-turn memory across just 167 lines).

Not for: Not suitable for analyses requiring empirical economic facts, data, or literature support (no external retrieval or verbatim quoting; relies solely on model deduction); not suitable for multi-school economic comparisons or policy feasibility evaluations (perspectives are strictly internal to Austrian economics without external rebuttals); not suitable for professional writing requiring authoritative citations, as character statements are simulated and quoting them constitutes apocryphal attribution; and not suitable for host environments that do not support parallel sub-agents or fail to recognize the `sonnet` model identifier (which may degrade into a single model roleplaying both personas or throw an error).

Install Agent-Ready

① Verified Mirror 2026-09-15

Option A · Direct DownloadDownload dbs-chatroom-austrian.tar.gz

sha256: 0c5cac0e4ef44304…

Option B · Copy Guide for Agent

Agents automatically download and verify mirrors from this JSON guide.

② Upstream GitHub

GitHub Repo ↗

Pinned to commit 8b8e33f1ec.