huawei-cloud-cloudrobo-dispatch
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes natural-language task descriptions and retrieves logs from external robot executions, which represents an indirect prompt injection surface.
- Ingestion points: User-provided
taskstrings increate-taskand external logs/results retrieved viashow-task-result(SKILL.md). - Boundary markers: None explicitly shown in the provided command templates.
- Capability inventory: The agent can execute system commands via the
cloudroboCLI. - Sanitization: The instructions explicitly mandate sanitizing task content and implement
validate_safe_idto prevent path traversal in identifier parameters (SKILL.md, references/verification-method.md). - [SAFE]: All network communications are directed towards official Huawei Cloud domains (*.myhuaweicloud.com) associated with the vendor.
- [SAFE]: The dependency
hw-cloudrobo-dispatchis a verified vendor-specific package. - [SAFE]: Authentication is correctly handled using environment variables for AK/SK credentials, and the documentation contains explicit warnings against hardcoding secrets.
Audit Metadata