Threat modeling is simply asking, before attackers do: what could go wrong, and what will we do about it? It is defensive by nature and pairs well with authorized scanning.
Start with assets
List what is worth protecting: user credentials and tokens, personal data, payment details, API keys shipped in the app, and the backend behind it. Everything else follows from protecting these.
Map the entry points
- the app's local storage, logs and cache
- traffic between app and API
- third-party SDKs and their permissions
- the backend's public endpoints
Apply STRIDE
Walk each entry point through Spoofing, Tampering, Repudiation, Information disclosure, Denial of service and Elevation of privilege. For each plausible threat, note an existing control or an action item.
Mobile-specific risks
- Insecure storage: secrets in
SharedPreferences/plist instead of the keystore/keychain. - Weak transport: missing certificate validation, or no TLS pinning on sensitive flows.
- Hardcoded secrets: API keys embedded in the binary.
- Over-broad permissions: SDKs requesting more than they need.
Where ZekaForge fits
The backend's web surface maps directly to ZekaForge Cyber AI's authorized checks — TLS configuration, security headers, OWASP categories and outdated components — while Device Cloud lets you observe the app's real runtime behavior on a device. Findings feed the AI assistant, which prioritizes them by risk.
The output
A short, living document: assets, the threats you found, the controls you have, and a ranked list of fixes. Revisit it whenever the app's attack surface changes.