investment-ai-product-judgmentlisted
Install: claude install-skill qihangzhang-272/agent-skill-library
# Investment AI Product Judgment
This node answers: does the AI product itself make sense?
It is the product judgment layer inside `domain-investment`.
## References
Always read:
- `references/ai-product-analyzer.md`
Read selectively:
- `references/business-model.md` for pricing, inference cost, revenue quality, subscription vs usage, margin, bundling, or business-model risk.
- `references/data-agent.md` for data agent, BI, Text-to-SQL, enterprise data workflow, context layer, or analytics agent products.
- `references/narrative-audit.md` for positioning, pitch deck, NOT positioning, narrative line, or old-paradigm packaging risk.
## Output
```text
AI-native product judgment:
Reference usage:
One-sentence positioning:
Problem validity:
Solution validity:
Product form:
Why now:
Market fit:
Business model:
Competition:
Traction:
Team / GTM:
Financial / ask:
Verdict: good case / bad case / watch
Strongest product argument:
Weakest product gap:
Narrative line:
Handoff to investment package:
```
## 完成标准
- 先判断并读取所有适用内部 reference；投资工作流至少读取一个适用 reference 并记录选择理由，输入过薄时不得降级为轻量意见
- 完整输出 Purpose、Problem、Solution、Product、Why Now、Market、Business Model、Competition、Traction、Team/GTM、Financials/Ask 十一段判断
- 分别检验 AI 原生性、真实业务价值和叙事强度，避免把技术新颖度直接等同投资价值
- 每个核心产品结论都连接输入事实、证据或明确假设，禁止无来源补造产品能力
- 给出最强项、最弱项、反方解释、结论边界以及会改变判断的新证据
- 经合理查找仍不可得的信息必须用业务语言写清缺失事项、不可得原因、对结论的影响、当前保守处理和重新核验条件后继续；不得将未知写成零或事实
- 形成可独立检查的产物：$artifact。
- 在 Mode 内调用与独立调用执行相同标准；无法满足的项目必须说明缺口与影响，不得伪造完成。