横纵分析法

Cross-Axis Analysis Framework. Use when you need to systematically research an unfamiliar domain, product, company, or technology concept through longitudina...

Security Scan

This is an instruction-only research framework (prompt template + usage steps) and its requested actions and resources are consistent with its stated purpose.

Purpose & Capability

The skill is a research framework that loads an internal prompt template and directs the agent to run longitudinal + cross-sectional analysis via LLMs. It declares no external credentials, binaries, or installs — nothing requested is disproportionate to doing deep/quick research.

Instruction Scope

SKILL.md instructs the agent to read the included references/prompt-template.md, select a version, fill in the research target, and send it to an LLM. It does not instruct reading unrelated files, accessing environment variables, or contacting unexpected endpoints. The only external interaction implied is sending prompts to LLM services (ChatGPT/Claude/Gemini), which is consistent with the skill's purpose.

Install Mechanism

No install spec and no code files — the skill is instruction-only. There is nothing to download or write to disk during installation.

Credentials

The skill requires no environment variables, credentials, or config paths. It does not request access to unrelated secrets or system resources.

Persistence & Privilege

always is false and there is no indication the skill modifies other skills or system-wide settings. Autonomous model invocation is allowed by default but is expected for this type of instruction-only research skill.

Assessment

This skill is a template-driven research method and appears internally consistent. Before using it, consider: (1) privacy — the research target text you fill into the prompt will be sent to whatever LLM the agent uses (OpenAI/Anthropic/Google/etc.), so avoid inserting confidential or sensitive material; (2) cost & latency — the 'deep' version asks for very long outputs (10k–30k words) which may be expensive and slow with some models; (3) factual verification — LLM outputs can hallucinate, so independently verify important claims; (4) autonomy settings — if you are concerned about the agent initiating long-running or costly LLM calls on its own, restrict autonomous invocation or monitor model usage. Other than those operational considerations, there are no hidden installs, requested credentials, or unrelated file accesses.

Cross-Axis Analysis

通过纵向时间深度 + 横向竞争格局的交汇分析，快速建立对陌生领域的系统性认知。

工作流程

1. 确认研究对象 — 明确分析目标（产品/公司/技术/人物/趋势）
2. 加载 Prompt 模板 — 读取 references/prompt-template.md
3. 选择版本：
   • 深度版：适合学术研究、投资决策、战略规划（10,000-30,000 字）
   • 快速版：适合紧急判断、初步了解（≤1,500 字）
4. 填充并执行 — 将 [研究领域/对象] 替换为实际目标，发送给 AI
5. 交付结果 — 将 AI 输出整理后交付用户，必要时补充自己的解读

何时用这个方法

场景	版本	预期产出
学术研究（文献综述）	深度版	完整发展脉络 + 竞品对比
选校/选公司决策	深度版	生态位分析 + 横纵交汇判断
投资分析（币/股）	深度版	路径依赖分析 + 未来展望
快速了解新领域	快速版	3 个转折点 + 3 个竞品
小说世界观调研	深度版	同类作品生态位 + 差异化机会

方法优势

• 结构清晰 — 纵+横+交汇，不遗漏关键维度
• 天然适配 AI — 把 AI 的研究能力装了方向盘
• 横纵交汇是灵魂 — 把"三年前决策"和"今天地位"连起来，产生真正洞察

局限性

• 冷门领域可能信息不足，需补充人工调研
• 深度版篇幅大，不适合需要即时决策的场景（用快速版）
• 分析质量取决于 AI 的信息检索能力