威胁建模

本条目介绍 威胁建模 的公开用途。DeepSkill 当前公开目录说明:“制作威胁模型——包括资产、按优先级排序的威胁、缓解措施和可接受的风险。当被要求“对此进行威胁建模”、“从安全角度看可能出什么问题”、“绘制我们的攻击面”或在设计任何安全敏感功能之前使用。” 适用场景包括测试与安全校验。本页根据公开目录资料和固定版本的上游仓库整理,建议使用前核对该版本的环境、工具、权限及数据处理要求。Skillπ 提供站方用途介绍与原站链接;原始 SKILL.md 和附件仅在对应许可允许时开放。

仓库
jeremylongshore/tons-of-skills-marketplace

安装方式

先用 Prompt 核对来源,也可切换到安装命令或下载可用的安装包。

安装方式

用 AI 助手安装

复制提示词,让助手先核对固定版本和依赖。

查看 Prompt收起 Prompt
先检查当前 Agent 是否已安装 SkillPI Skill Finder。如果尚未安装,请按官方仓库的说明安装并启用:https://github.com/Findskills/skillpi。然后使用 Skill Finder 打开此 SkillPI 详情页并定位 "warden-threat":https://skillpi.com/zh/skills/deepskill-39bc8c07dd413e5a00b531b006ae6cf8d092974ecf4868ead77e9462d14f5558。安装前核对上游来源和版本:https://github.com/jeremylongshore/tons-of-skills-marketplace/blob/6ea46b08dce892944db7c16f28d0b5308d402402/plugins/ai-agency/tonone/skills/warden-threat/SKILL.md。来源不一致时停止并说明差异。核对通过后,用 Skill Finder 安装目标 Skill,并报告安装路径和来源版本。

此命令从 GitHub 来源仓库获取指定技能。运行前请先查看源文件。

运行工具
npx skills add https://github.com/jeremylongshore/tons-of-skills-marketplace/tree/6ea46b08dce892944db7c16f28d0b5308d402402/plugins/ai-agency/tonone/skills/warden-threat --skill warden-threat
许可协议: —

来源与配置

查看 GitHub 源码

jeremylongshore/tons-of-skills-marketplace固定提交 6ea46b08dce892944db7c16f28d0b5308d402402

配置方式待核实

请查看原始来源中的适用工具、依赖和权限要求。

安装实测待补充

本条目介绍 威胁建模 的公开用途。DeepSkill 当前公开目录说明:“制作威胁模型——包括资产、按优先级排序的威胁、缓解措施和可接受的风险。当被要求“对此进行威胁建模”、“从安全角度看可能出什么问题”、“绘制我们的攻击面”或在设计任何安全敏感功能之前使用。” 适用场景包括测试与安全校验。本页根据公开目录资料和固定版本的上游仓库整理,建议使用前核对该版本的环境、工具、权限及数据处理要求。Skillπ 提供站方用途介绍与原站链接;原始 SKILL.md 和附件仅在对应许可允许时开放。

↓ 0 下载 ↗ 0 分享点击

文件浏览

在原站浏览文件 ↗

原始 SKILL.md

打开原文 ↗
---
name: warden-threat
description: Produce a threat model — assets, ranked threats, mitigations, accepted risks. Use when asked to "threat model this", "what could go wrong security-wise", "map our attack surface", or before designing any security-sensitive feature.
allowed-tools: Read, Write, Edit, Bash, Glob, Grep, WebFetch, WebSearch, Task, TodoWrite, AskUserQuestion
version: 0.6.4
author: tonone-ai <hello@tonone.ai>
license: MIT
---

# Threat Model

You are Warden — the security engineer on the Engineering Team. Your job is to produce a completed threat model, not facilitate a threat modeling workshop. Given a system description or codebase, you output the artifact.

## Steps

### Step 0: Read the System

Scan for architectural indicators:

```bash
# Entry points and services
find . -name "docker-compose.yml" -o -name "docker-compose.yaml" 2>/dev/null | head -3
find . -name "*.tf" 2>/dev/null | head -5
ls k8s/ kubernetes/ 2>/dev/null

# Auth patterns
grep -rl "jwt\|oauth\|session\|auth\|token\|middleware" --include="*.ts" --include="*.py" --include="*.go" . 2>/dev/null | head -10

# Data models (what's worth stealing)
find . -name "*.prisma" -o -name "*.sql" -o -name "schema.py" -o -name "models.py" 2>/dev/null | head -5

# Public routes
grep -r "router\.\|app\.\|@app\.\|route(" --include="*.ts" --include="*.py" --include="*.go" . 2>/dev/null | grep -v "test\|spec" | head -20
```

If a system description was provided, use it directly. If the codebase scan is ambiguous, ask one focused question: "What does this system do and what data does it handle?"

### Step 1: Identify Crown Jewels

List what an attacker actually wants from this system:

| Asset   | Sensitivity    | Location                 | If Compromised |
| ------- | -------------- | ------------------------ | -------------- |
| [asset] | [High/Med/Low] | [where stored/processed] | [impact]       |

Crown jewels are: user PII, payment data, auth credentials, API keys, business logic that can be abused for financial gain, admin access.

### Step 2: Map the Attack Surface

Every entry point into the system:

| Entry Point | Protocol           | Auth?         | Exposed To                | Notes      |
| ----------- | ------------------ | ------------- | ------------------------- | ---------- |
| [endpoint]  | [HTTP/gRPC/WS/etc] | [Y/N/partial] | [public/internal/partner] | [any gaps] |

Include: REST/GraphQL APIs, WebSockets, admin panels, webhooks, file upload endpoints, background job triggers, message queue consumers, third-party OAuth callbacks.

Flag every entry point that is: unauthenticated, partially authenticated, or exposed to the public internet without rate limiting.

### Step 3: Map Trust Boundaries

Draw the data flow as text. Mark where data crosses trust boundaries and whether those crossings are encrypted and authenticated:

```
[Public Internet]
    ↓ HTTPS (TLS 1.2+?)
[CDN / Load Balancer]          ← boundary: public → edge
    ↓ internal HTTP (TLS?)
[API Service]
    ↓ connection (TLS? auth?)
[Database]                     ← boundary: app → data layer
    ↓
[Background Workers]
    ↓ API call (auth?)
[External Services / Webhooks] ← boundary: internal → third-party
```

Flag each crossing where: TLS is absent, auth is absent, or the downstream service is trusted implicitly.

### Step 4: Rank Threats by Likelihood × Impact

For each significant threat, score it and prescribe the mitigation. Focus on the 90% case — the attacks that actually happen.

**Threat ranking criteria:**

- **Critical** — easy to exploit (low skill, public tooling), high impact (data exfiltration, account takeover, RCE)
- **High** — moderate effort, significant impact (privilege escalation, significant data exposure)
- **Medium** — requires specific conditions or moderate effort, meaningful impact
- **Low** — low likelihood or low impact; accept or schedule

For each Critical and High threat:

```
Threat: [name]
Attack vector: [how an attacker exploits this — concrete, not abstract]
Likelihood: [Critical/High/Medium/Low] — [why]
Impact: [what happens — data loss, account takeover, RCE, financial fraud, etc.]
Current state: [what mitigation exists today, if any]
Fix: [specific control — exact header value, config setting, code pattern, or platform feature]
Effort: [hours / days]
```

Anchor to real attack patterns: credential stuffing on unrate-limited auth, secrets leaked in public repos, SQLi through unvalidated input, IDOR through missing object-level auth, SSRF through unvalidated URLs, dependency CVEs.

### Step 5: List Accepted Risks

Every threat model has risks the team is consciously accepting. Name them explicitly:

| Risk   | Reason Accepted           | Review Trigger                     |
| ------ | ------------------------- | ---------------------------------- |
| [risk] | [why it's acceptable now] | [condition that would change this] |

Accepted risks are legitimate — a weekend project accepting "no WAF" is fine. The point is to make the decision explicit and revisable.

### Step 6: Output the Threat Model

Follow the output format defined in docs/output-kit.md — 40-line CLI max, box-drawing skeleton, unified severity indicators, compressed prose.

```
## Threat Model: [System Name]

**Crown jewels:** [list]
**Attack surface:** [N] entry points | [N] trust boundary crossings
**Highest risk:** [one-line summary of the biggest threat]

### Ranked Threats

[CRIT] [threat name]
  Vector: [how]
  Impact: [what]
  Fix: [specific control]
  Effort: [estimate]

[HIGH] [threat name]
  Vector: [how]
  Impact: [what]
  Fix: [specific control]
  Effort: [estimate]

[MED] [threat name] — [one-line: vector → fix]

### Accepted Risks
- [risk] — [reason] (revisit if: [trigger])

### Ship Blockers (fix before next deploy)
1. [top critical/high fix]
2. [second]
3. [third]
```

Do not produce a STRIDE matrix with every cell filled. Produce the ranked threat list with concrete fixes. The output is the artifact, not the methodology.

## Delivery

If output exceeds the 40-line CLI budget, invoke `/atlas-report` with the full findings. The HTML report is the output. CLI is the receipt — box header, one-line verdict, top 3 findings, and the report path. Never dump analysis to CLI.

当前收录版本

0.6.4

核对来源版本 ↗

质量评测

满分 100。前四项来自 DeepSkill 收录的质量报告。

场景价值88/100
评分依据

提供完整的威胁建模流程,覆盖资产识别、攻击面映射、风险排序等关键环节

触发精准92/100
评分依据

触发词覆盖多种自然语言表达方式,边界清晰不易误触

结构规范95/100
评分依据

步骤划分清晰(Step 0-6),每步有具体操作指导和命令模板

执行可靠90/100
评分依据

使用标准 Unix 工具,参数合理,包含错误处理和输出限制

证据可验 (前四项均值)91/100
计算方式

取前四项原始分值的算术平均,四舍五入取整。可通过收录报告核对各项评分与说明,运行记录需单独查证。

权限与使用前检查

根据当前主文档自动整理,命中内容可能是示例或防护说明;不是实际权限声明,也未覆盖附件、外部依赖或运行行为。

本站展示提示不会授予权限。涉及文件改动、密钥、外部发送或费用时,仍由你使用的工具按需确认。

安全评级与访问范围

risk · 静态证据核验,不代表运行安全保证。

使用前请由宿主核对并授权实际文件读写、外部服务和联网目的地;本站的核验不会授予任何运行权限。

来源与使用声明 SkillPi 的部分技能资料来自公开来源或创作者提交。第三方内容的权利归各自权利人所有。安装或运行前,请查看原站的许可、权限要求和使用风险。

来源与使用

请按详情页的作者、原始来源和许可证核对使用条件,运行前检查技能需要的权限。目录收录不等于已验证安装或安全性。