技能详情（站内镜像，无评论）
许可证：MIT-0
MIT-0 ·免费使用、修改和重新分发。无需归因。
版本：v1.1.0
统计：⭐ 0 · 110 · 0 current installs · 0 all-time installs
⭐ 0
安装量（当前） 0
🛡 VirusTotal ：良性 · OpenClaw ：良性
Package：agisearch/first
安全扫描（ClawHub）
- VirusTotal ：良性
- OpenClaw ：良性
OpenClaw 评估
The skill's code and instructions match its stated purpose (local first-principles reasoning); it stores cases locally under your home directory, does not request credentials, and has no external network or download behavior.
目的
The name/description (first-principles reasoning) align with the provided Python scripts: reasoning heuristics, case creation, scoring, listing, exporting, and local pattern promotion. The files implement the declared functionality and there are no unrelated dependencies or unexplained external-service credentials.
说明范围
The SKILL.md is scoped to reasoning tasks and does not instruct any network calls or access to unrelated system paths. However, the included scripts persist data (cases, patterns, exports) to disk; SKILL.md calls this a 'local-first' skill but does not explicitly enumerate the storage behavior (see storage path below). The skill does not read other user config or secrets.
安装机制
There is no install spec or remote download. All code is packaged with the skill; nothing will be pulled from external URLs during install.
证书
The skill requests no environment variables or credentials. Its only privileged action is writing files under the user's home directory (no access to system-level credentials or unrelated service tokens).
持久
The skill creates and updates files under ~/.openclaw/workspace/memory/firstprinciples (cases.json, patterns.md, case_index.md, exports). always is false and autonomous invocation is allowed (platform default). Persisting user-provided problem text locally is expected behavior but is important to note for sensitive data.
综合结论
This skill appears coherent and self-contained. Before enabling: (1) be aware it will create and persist data under ~/.openclaw/workspace/memory/firstprinciples (cases.json, patterns.md, case_index.md, and an exports folder); do not submit sensitive secrets or personally identifying content as problem text if you don't want it stored. (2) If desired, inspect the packaged Python files (scripts/lib/storage.py and scripts/lib/reasoning.py) to con…
安装（复制给龙虾 AI）
将下方整段复制到龙虾中文库对话中，由龙虾按 SKILL.md 完成安装。
请把本段交给龙虾中文库（龙虾 AI）执行：为本机安装 OpenClaw 技能「Firstprinciples thinking」。简介：Think from first principles instead of analogy. Break problems into fundamental…。
请 fetch 以下地址读取 SKILL.md 并按文档完成安装：https://raw.githubusercontent.com/openclaw/skills/refs/heads/main/skills/agisearch/first/SKILL.md
（来源：yingzhi8.cn 技能库）SKILL.md
---
name: Firstprinciples thinking
slug: first
description: Think from first principles instead of analogy. Break problems into fundamental truths, challenge assumptions, and rebuild better solutions from scratch.
---
# Firstprinciples thinking
A local-first reasoning skill for breaking complex problems down to their most fundamental truths and rebuilding solutions from the ground up.
Use this when the user:
- is stuck in conventional wisdom
- wants original thinking instead of recycled advice
- needs to separate facts from assumptions
- wants to redesign a strategy, product, business model, or decision from scratch
- asks whether something is truly necessary, optimal, or possible
This skill should:
1. Clarify the real objective.
2. Surface explicit and hidden assumptions.
3. Separate facts from analogies, conventions, and defaults.
4. Break the problem into irreducible components.
5. Identify real constraints, variables, and tradeoffs.
6. Reconstruct a better solution from the ground up.
7. Deliver a practical answer with clear next actions.
## Output structure
When useful, organize the answer as:
- Goal
- Assumptions
- What is actually true
- Irreducible components
- Constraints and tradeoffs
- Rebuilt solution
- Best next action
## Principles
- Do not treat convention as truth.
- Do not treat common practice as proof.
- Do not treat analogy as understanding.
- Start from what must be true.
- Rebuild toward something usable.
- Stay grounded in the user's real goal.
## Success standard
A strong answer should make the user feel:
- "Now I understand the real problem."
- "These assumptions were weaker than I thought."
- "This solution is more fundamental and more original."
- "I can actually act on this."