usb-debugging
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is primarily instructional, guiding the agent through hardware diagnosis workflows without executing dangerous commands or requesting excessive permissions.
- [SAFE]: Includes a proactive security instruction to redact sensitive information (credentials, private serial numbers) from logs and reports before they are shared.
- [SAFE]: External links target official documentation sites for TinyUSB, Zephyr Project, MicroPython, and Microsoft, which are trusted and well-known resources in the embedded development community.
- [INDIRECT_PROMPT_INJECTION]: The skill provides a framework for analyzing untrusted data such as USB descriptors and system logs.
- Ingestion points: System logs, USB descriptors, and hardware identity strings provided by the user (SKILL.md, references/workflow.md).
- Boundary markers: None explicitly defined, though the instructions provide a conceptual framework for separating physical evidence from application data.
- Capability inventory: The skill does not define or use tools for network exfiltration, file writing, or arbitrary command execution.
- Sanitization: Instructions include a specific requirement to redact sensitive data from reports.
Audit Metadata