threat-modeling

pvillega 的头像作者 pvillega

在实施身份验证、文件上传、支付或外部API时使用。系统应用STRIDE框架。触发器:"身份验证"、"文件上传"、"支付"、"多租户"、"外部API"。如果认为"我了解安全" - 仍然使用它。

📖 中文导读:在实施身份验证、文件上传、支付或外部API时使用。系统应用STRIDE框架。触发器:"身份验证"、"文件上传"、"支付"、"多租户"、"外部API"。如果认为"我了解安全" - 仍然使用它。 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。

name: threat-modeling description: Use when implementing auth, file uploads, payments, or external APIs. Applies STRIDE framework systematically. Triggers: "authentication", "file upload", "payment", "multi-tenant", "external API". If thinking "I know security" - use this anyway.

Threat Modeling

MANDATORY FIRST STEP

TodoWrite: Create 18+ items (6 STRIDE categories × 3 items each)

  • Define assets/entry points
  • Enumerate threats per STRIDE category
  • Risk scoring
  • Identify controls
  • Document threat model
  • Verification checkpoint

Quick-Start Templates

Authentication Feature

Assets: User credentials, session tokens, reset tokens Entry points: Login, signup, password reset endpoints

STRIDE Focus:

Threat Check
Spoofing Credential stuffing protection? Rate limiting?
Tampering Token manipulation? Secure cookies?
Repudiation Login attempt logging?
Info Disclosure Token leakage? Error messages reveal info?
DoS Rate limiting? Account lockout?
EoP Session hijacking? Privilege escalation?

File Upload Feature

Assets: Stored files, file metadata, server filesystem Entry points: Upload endpoint, file retrieval

STRIDE Focus:

Threat Check
Spoofing Impersonate uploader?
Tampering Malicious files? Path traversal?
Repudiation Upload logging?
Info Disclosure Access others' files?
DoS Large file attacks?
EoP Execute uploaded code?

Multi-Tenant API

Assets: Tenant data, API endpoints, configurations Entry points: All endpoints with tenant context

STRIDE Focus:

Threat Check
Info Disclosure Cross-tenant data leakage?
EoP Access other tenant's data?
Tampering Modify tenant_id in requests?
Spoofing Impersonate other tenant?

STRIDE Process

Apply each category systematically (S→T→R→I→D→E):

  1. Spoofing: Can attacker impersonate?
  2. Tampering: Can attacker modify data/code?
  3. Repudiation: Can attacker deny actions?
  4. Information Disclosure: Can attacker access secrets?
  5. Denial of Service: Can attacker disrupt?
  6. Elevation of Privilege: Can attacker gain unauthorized access?

Risk Scoring: Likelihood (1-5) × Impact (1-5)

  • 15-25: Critical (address immediately)
  • 6-14: High (before release)
  • 1-5: Low (track)

Verification Checkpoint

Before marking complete, verify 3 specifics:

  1. ✅ Each STRIDE category has ≥2 concrete threats identified
  2. ✅ All high-risk (score ≥6) threats have controls documented
  3. ✅ Threat model includes data flow diagram with trust boundaries

Response Templates

"This is over-engineering for a simple feature"

Security bugs cost 30x more post-release than during design. STRIDE takes 15-30 minutes and prevents deployment blockers. Which STRIDE categories don't apply to this feature?


Red Flags

Thought Reality
"Ad-hoc brainstorming is faster" Misses 60% of threats
"Everything is critical" No risk prioritization = wrong fixes
"We'll security review later" 30x more expensive post-implementation