uniapp-project
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of technical documentation and Vue code snippets for developers using uni-app. Analysis of the instructions and examples revealed no malicious patterns.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill uses safe placeholders (e.g., 'https://api.example.com') for network operations. Examples demonstrating authorization headers or storage operations use dummy data (e.g., 'token123').
- [INDIRECT_PROMPT_INJECTION]: The skill documents components that ingest user input and APIs with side effects (network, storage, filesystem), which are inherent to the uni-app technical stack.
- Ingestion points: User input components such as input, textarea, and uni-combox in files like 'examples/components/built-in/input.md' and 'examples/uni-ui/uni-combox.vue'.
- Boundary markers: Not applicable, as the files provide reference code for copy-pasting rather than runtime logic for the agent.
- Capability inventory: Network requests (uni.request), local storage (uni.setStorage), and file system access (uni.saveFile) are documented as part of the framework's capabilities.
- Sanitization: Standard for example code, focusing on basic functionality rather than complex validation logic.
Audit Metadata