huawei-cloud-mrs-host-alarm-diagnose
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The Python API client script (
scripts/lakewatch_api_client.py) invokes a local security binary (CryptoAPI) via thesubprocessmodule to handle encryption and decryption tasks on Linux platforms. The skill's knowledge base also provides documentation for various system-level diagnostic commands (e.g.,smartctl,lsblk,gsql) intended for manual execution by the operator during deep investigation. - [EXTERNAL_DOWNLOADS]: The skill requires the installation of the
pyyamlandcryptographyPython libraries from standard package registries to facilitate configuration parsing and local AES encryption on Windows systems. - [PROMPT_INJECTION]: The diagnostic workflow involves ingesting external log data via the LakeWatch API, which presents an attack surface for indirect prompt injection. This risk is managed through the use of keyword-based filtering and explicit instructions for the agent to summarize log content rather than directly obeying any instructions contained within it.
- [SAFE]: The skill implements robust credential management by requiring service account passwords to be encrypted via a dedicated command before storage. Additionally, cached authentication tokens are stored with restricted file system permissions (
0600) to prevent access by other users on the system.
Audit Metadata