# Sm Model Check

> 证券分析师专用的财务模型检查 skill。用于审阅盈利预测与估值模型的关键假设、勾稽关系、收入利润桥、敏感性、情景假设和潜在断裂点，帮助发现模型里最容易误导投资结论的地方。

- **Type:** Skill
- **Install:** `agentstack add skill-joansongjr-investor-harness-sm-model-check`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [joansongjr](https://agentstack.voostack.com/s/joansongjr)
- **Installs:** 0
- **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [joansongjr](https://github.com/joansongjr)
- **Source:** https://github.com/joansongjr/investor-harness/tree/main/skills/sm-model-check

## Install

```sh
agentstack add skill-joansongjr-investor-harness-sm-model-check
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# SM Model Check

这个 skill 用于检查模型有没有把研究结论"算歪"，重点不是排版，而是看模型逻辑是否站得住。

## 强制流程（v0.3 硬约束）

> ⛔ **任何分析输出之前**，必须严格执行 [`../../core/preamble.md`](../../core/preamble.md) 的 5 步开始前流程
>
> ⛔ **任何输出完成之前**，必须严格执行 [`../../core/postamble.md`](../../core/postamble.md) 的 6 步结束后流程
>
> 输出归档按 [`../../core/output-archive.md`](../../core/output-archive.md) 命名规范
> 输出验收按 [`../../core/acceptance.md`](../../core/acceptance.md) 清单逐条自检
>
> **跳过任何一环视为未完成任务。**

Model Check 特别注意：preamble Step 4 必须包含对比基线（模型输出 vs 历史财务 vs 一致预期）。若用户未提供模型文件，可以先按研究逻辑层面做"模型框架检查"，但必须在 [Preflight] 中明确声明。

适用场景：

- 审阅盈利预测模型
- 更新财报后的模型
- 检查估值假设
- 给基金经理或研究总监过模型前做自检

## 核心任务

- 检查关键假设是否清晰
- 检查收入、毛利、费用、现金流之间是否勾稽
- 检查驱动变量与投资逻辑是否一致
- 检查估值方法和情景假设是否自洽
- 找出最可能误导结论的脆弱点

## 输出格式

- `模型目的`
- `核心假设清单`
- `收入驱动检查`
- `利润率 / 费用率检查`
- `现金流与资产负债表检查`
- `估值假设检查`
- `敏感性最高的三个变量`
- `最容易误导结论的地方`
- `需要立刻回头核验的项目`

## 检查重点

- 收入增长是否能拆到量、价、结构或客户层面
- 毛利率变化是否有合理驱动
- 费用率变化是否和扩张节奏匹配
- 现金流是否支持利润表结论
- Capex、折旧、营运资本是否自洽
- 估值倍数或目标价假设是否与赛道、阶段、可比公司匹配

## 约束

- 不要只说"模型合理"或"需要进一步验证"
- 必须指出哪一行逻辑最脆弱，以及为什么
- 若用户未提供模型文件，也可以先按研究逻辑层面做"模型框架检查"

## 参考

- [../../core/evidence.md](../../core/evidence.md)
- [../../core/compliance.md](../../core/compliance.md)
- [../../core/templates.md](../../core/templates.md)
- [../../core/adapters.md](../../core/adapters.md)

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [joansongjr](https://github.com/joansongjr)
- **Source:** [joansongjr/investor-harness](https://github.com/joansongjr/investor-harness)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-joansongjr-investor-harness-sm-model-check
- Seller: https://agentstack.voostack.com/s/joansongjr
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.