appstore-review
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and analyze untrusted data from the user's project files, such as Swift source code and configuration files.\n
- Ingestion points: The agent reads
.swiftfiles,Info.plist,PrivacyInfo.xcprivacy, entitlements, and asset catalogs (SKILL.md).\n - Boundary markers: No explicit markers or instructions are provided to the agent to distinguish between data and potential instructions within the analyzed files.\n
- Capability inventory: The agent's actions are limited to inspecting file contents and generating a report; it is not instructed to perform any destructive shell commands or external network writes.\n
- Sanitization: There is no explicit sanitization or filtering logic applied to the content read from the project files.\n- [EXTERNAL_DOWNLOADS]: The documentation references an installation script and a repository on GitHub (github.com/3paws-ai/mobile-ai-skills.git). These are established as legitimate vendor resources from the skill's author.\n- [COMMAND_EXECUTION]: The skill requires the agent to perform local filesystem searches and file inspections (e.g., searching for 'Lorem Ipsum' or 'TODO' markers) to fulfill its auditing purpose. This execution is limited to the scope of the local project.
Audit Metadata