ios-developer
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions proactively recommend secure coding practices, including using Keychain Services for sensitive data storage and implementing App Transport Security (ATS) for network communications.
- [SAFE]: References to external libraries and tools (e.g., Swinject, SDWebImage, Kingfisher, Fastlane, Sentry) are standard, well-known resources within the iOS development ecosystem.
- [INDIRECT_PROMPT_INJECTION]: The skill includes instructions to process external feature briefs, which represents a potential injection surface.
- Ingestion points: The
resources/implementation-playbook.mdfile specifies "feature brief" as a required input for the development workflow. - Boundary markers: The instructions do not define specific delimiters or "ignore previous instructions" guards for the external content.
- Capability inventory: The skill is capable of generating Swift code, UI components, and networking logic. It does not have access to automated file system modification or external network exfiltration tools in its current configuration.
- Sanitization: No explicit sanitization steps for external inputs are provided in the implementation guidelines.
Audit Metadata