huawei-cloud-cdn-traffic-anomaly-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFE
Full Analysis
- [CREDENTIALS_SAFE]: The skill provides explicit instructions to avoid hardcoding or providing Access Keys (AK) and Secret Keys (SK) in the conversation. It directs users to use standard, secure methods like environment variables or the
hcloud configureinteractive setup. - [COMMAND_EXECUTION]: The skill executes
hcloud(KooCLI) commands and a local Python script (scripts/cdn_timestamp.py). All commands are read-only (e.g.,ListDomains,ShowDomainStats) and target the vendor's own infrastructure. The skill includes a 'Prohibited Operations' section that explicitly forbids write or delete operations. - [DATA_EXFILTRATION]: No patterns of data exfiltration were detected. Network operations are limited to communicating with the official Huawei Cloud API endpoints via the
hcloudCLI tool for the primary purpose of traffic analysis. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface as it processes external data from CLI tool outputs. However, it implements a validation step to ensure the target domain exists in the user's account and uses specific, well-defined CLI parameters, which minimizes the risk of schema confusion or unexpected command injection via tool output.
- [NO_CODE]: The skill includes one Python script (
scripts/cdn_timestamp.py) which was audited and found to contain only benign time calculation logic using standard libraries (argparse,datetime). It does not useeval(),exec(), or any network-related modules.
Audit Metadata