huawei-cloud-cost-estimation
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill interacts with the Huawei Cloud environment using the
hcloudCLI. It maintains a high level of safety by requiring a mandatory--dryrunexecution for all provisioning commands and restricting the agent's write capabilities to a predefined allowlist of 74 specific operations. - [CREDENTIALS_UNSAFE]: The skill implements a secure credential handling policy by explicitly instructing the agent to refuse any API keys or tokens provided in the chat interface, directing users to use local CLI configurations instead.
- [EXTERNAL_DOWNLOADS]: The documentation includes instructions for users to install the official Huawei Cloud CLI from vendor-controlled domains (
myhuaweicloud.com). These resources are documented for manual user setup and are not invoked autonomously by the agent. - [PROMPT_INJECTION]: The skill contains well-defined behavioral constraints that serve as internal guardrails, preventing the agent from bypassing pricing checks, confirmation steps, or safety filters regarding resource unsubscription.
Audit Metadata