Zum Inhalt springen
TakoAPI
Zurück zu den Skills

threat-modeling

von brandonwise

Coding Agents & IDEs
461Downloadzahlen laut clawskills.sh
0Aufrufe
0 LikesMelde dich an, um diesen Skill zu liken
—Noch nicht bewertet
Deine Bewertung

Kurzbeschreibung

Expert in threat modeling methodologies, security architecture review, and risk assessment using STRIDE, PASTA.

Beschreibung

Was dieser Skill macht

Structured threat modeling using STRIDE, PASTA, attack trees, and DREAD/CVSS scoring to identify and prioritize security risks in system designs. Covers application, network, infrastructure, and human attack layers. Produces documented threat models with mitigations and a record of accepted residual risks.

Automated scanners find known vulnerabilities but miss trust boundary violations, insider threats, and logic flaws that structured threat modeling with STRIDE and attack trees catches.

Wann du ihn einsetzt

  • Reviewing a new microservices architecture before deployment
  • Preparing security documentation for an upcoming audit
  • Identifying attack vectors in a third-party API integration
  • Prioritizing which vulnerabilities to remediate first
  • Running a threat modeling session with a development team

Beispiel-Workflow

EINGABE

User asks: Can you threat model our new customer payment API?

AGENT
  1. 1

    Define scope, assets, and trust boundaries for the payment API

  2. 2

    Generate a data flow diagram covering user, API gateway, backend, and payment processor

  3. 3

    Apply STRIDE analysis across each component and data flow

  4. 4

    Build attack trees for the highest-risk goals such as credential theft and transaction tampering

  5. 5

    Score threats using DREAD and produce a prioritized mitigation table

AUSGABE

A structured threat model document with a DFD, completed STRIDE table, ranked threat list with DREAD scores, recommended controls, and documented residual risks

Installation

clawhub install brandonwise/threat-modeling

Führe den CLI-Befehl aus oder füge den Prompt in Claude Code, Cursor, Codex oder deinen Coding-Agenten ein.

Links