oneos-wave-router
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional and serves as a router to other specialized skills within the OneOS ecosystem. It explicitly defines a boundary policy of 'only guiding, not acting' (只指路，不办事), preventing it from making unauthorized changes to code or documentation.
- [SAFE]: No network operations, sensitive file access, or command execution patterns were found in any of the analyzed files.
- [SAFE]: No obfuscation, prompt injection, or persistence mechanisms are present. The persona instructions are standard role-play for agent personality and do not attempt to bypass safety filters.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user queries to determine the correct target skill. However, because the skill has no dangerous capabilities (no file writes, network calls, or code execution), the risk is negligible.
- Ingestion points: User query input defined in
SKILL.md. - Boundary markers: None explicitly defined for user input.
- Capability inventory: No subprocess calls, exec/eval, file-write, or network operations found across any scripts.
- Sanitization: None detected.
Audit Metadata