← 返回 Skills 市场
48h-Expert-Methodology
作者
Lawliet-ai
· GitHub ↗
· v1.0.1
· MIT-0
280
总下载
0
收藏
2
当前安装
2
版本数
在 OpenClaw 中安装
/install 48h-expert
功能描述
A meta-learning method compressing deep expertise into 48 hours by extracting core mental models, expert debates, and critical assessment questions for mastery.
使用说明 (SKILL.md)
Skill: 48h-Expert-Protocol (Cognitive-Compressor V2.1)
1. Core Assertion
System SHALL NOT output unstructured prose. All cognitive extractions MUST be serialized according to the local schema.json to ensure cross-skill interoperability.
2. Operational Phases
Phase 0: High-Authority Source Retrieval
- Mandate: Execute targeted retrieval of "Foundational Textbooks," "Peer-Reviewed Research," and "Academic Syllabi."
- Filtering: Prioritize
.edu,.gov, and high-impact industry white papers.
Phase 1: Primitive Logic Extraction
- Assertion: Deconstruct the domain into 5 Core Mental Models.
- Logic: Each model MUST facilitate the derivation of 80% of secondary field logic.
Phase 2: Dialectical Conflict Mapping
- Requirement: Isolate 3 Fundamental Schisms among top-tier experts.
- Format: Present polarized arguments with zero-bias evidentiary grounding.
Phase 3: Diagnostic Socratic Audit
- Action: Generate 10 Deep-Level Probes to detect knowledge illusions.
Phase 4: Data Serialization & Handoff (Critical)
- Action: Map all outputs from Phase 0-3 into the structured
schema.jsonformat. - Integrity Check: The resulting JSON MUST pass structural validation.
- Persistence: Write the final JSON to
~/.openclaw/swarm_tmp/expert_output.json.
3. Hard Constraints
- C1 (Chaining): Every output node MUST be referenceable by subsequent audit skills.
- C2 (Schema Compliance): Any deviation from
schema.jsonSHALL trigger a mandatory re-formatting cycle. - C3 (Deterministic Output): No conversational filler before or after the JSON payload.
安全使用建议
This skill appears to do what it says: fetch authoritative sources, extract core mental models, and write a validated JSON file. Before installing, consider: 1) Confirm you are comfortable with the agent performing web retrievals (it may crawl or fetch many pages); ensure your environment/network policy allows this. 2) The skill will write output to ~/.openclaw/swarm_tmp/expert_output.json but did not declare that config path — decide whether creating that file is acceptable and where you want artifacts stored. 3) The skill gives no guidance on handling paywalled or copyrighted sources, and it does not mention respecting robots.txt or rate limits; if scraping is a concern, restrict or sandbox the agent. 4) Because the skill forbids unstructured prose and insists on deterministic JSON, you may lose helpful narrative explanations; verify this behavior matches your needs. 5) If you want stronger assurance, run the skill in a sandboxed environment first, inspect the generated JSON and any fetched URLs, and confirm no unexpected network endpoints are contacted.
功能分析
Type: OpenClaw Skill
Name: 48h-expert
Version: 1.0.1
The skill bundle is a structured framework designed to guide an AI agent through academic research and knowledge synthesis. The instructions in SKILL.md focus on data retrieval, cognitive modeling, and serialization into a specific JSON format defined in schema.json. The only system interaction is the persistence of output to a local temporary directory (~/.openclaw/swarm_tmp/expert_output.json), which is consistent with the stated goal of cross-skill interoperability and does not exhibit signs of malicious intent, data exfiltration, or unauthorized system access.
能力评估
Purpose & Capability
The name/description (compress expertise into 48 hours) matches the runtime instructions: retrieve authoritative sources, extract mental models, map conflicts, generate diagnostic probes, and serialize into schema.json. There are no unrelated env vars, binaries, or credentials requested.
Instruction Scope
SKILL.md explicitly directs web retrieval of textbooks, peer-reviewed research, and syllabi and requires producing precisely structured JSON. This is in-scope for the task, but the instructions do not constrain how retrieval is performed (e.g., no guidance about respecting robots.txt, rate limits, or handling paywalled content), and they forbid any unstructured prose output which may limit human-readable justification. The skill also mandates prioritizing .edu/.gov sources but doesn't describe citation, licensing, or attribution policies.
Install Mechanism
Instruction-only skill with no install spec and no code files — lowest installation risk. Nothing is downloaded or written by an installer step.
Credentials
No environment variables, credentials, or special binaries are requested (proportional). However, the runtime instructions require writing output to a specific path in the user's home directory (~/.openclaw/swarm_tmp/expert_output.json) even though no required config paths were declared — a minor inconsistency that should be documented and approved by the user.
Persistence & Privilege
The skill is not always-enabled and can be invoked normally. It does request persistence by writing a JSON file to the user's home directory; this is reasonable for the stated purpose but is a persistent artifact created without the skill declaring config-path requirements. It does not request system-wide or other-skills modifications.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install 48h-expert - 安装完成后，直接呼叫该 Skill 的名称或使用
/48h-expert触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.1
- Introduced a strictly structured output format; all results must be serialized to the new `schema.json`.
- Added a new phase: Data serialization and validation, with outputs written to `~/.openclaw/swarm_tmp/expert_output.json`.
- Expanded source requirements to include government and industry white papers.
- Enforced several hard constraints: referenceability, strict schema validation, and deterministic output with no extraneous text.
- Replaced unstructured analysis with schema-driven, interoperable data for more robust downstream integration.
v1.0.0
This is a deep research plugin based on the strategy of "Meta-Learning." It forces AI to abandon superficial "summaries" and "explanations," instead focusing on 5 core thinking models, 3 fundamental points of divergence, and 10 expert-level stress test questions in the field. It is suitable for professionals who need to quickly enter new industries or new tracks.
元数据
常见问题
48h-Expert-Methodology 是什么？
A meta-learning method compressing deep expertise into 48 hours by extracting core mental models, expert debates, and critical assessment questions for mastery. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 280 次。
如何安装 48h-Expert-Methodology？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install 48h-expert」即可一键安装，无需额外配置。
48h-Expert-Methodology 是免费的吗？
是的，48h-Expert-Methodology 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
48h-Expert-Methodology 支持哪些平台？
48h-Expert-Methodology 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 48h-Expert-Methodology？
由 Lawliet-ai（@lawliet-ai）开发并维护，当前版本 v1.0.1。
推荐 Skills