huawei-cloud-smn-dms-message

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation and prerequisites include downloading the official Huawei Cloud CLI (KooCLI) installation script from an official vendor domain (cn-north-4-hcli.obs.cn-north-4.myhuaweicloud.com). This is a standard and expected operation for the skill's intended purpose.
  • [COMMAND_EXECUTION]: The Python dispatcher script (scripts/smn_dms_skill.py) executes the hcloud binary to perform cloud operations. It uses the secure subprocess.run method with argument lists, effectively preventing shell injection vulnerabilities.
  • [DYNAMIC_EXECUTION]: The skill dynamically constructs command-line arguments for the hcloud tool based on user input and pre-defined action templates. This is restricted to specific cloud operations and uses standard argument parsing to maintain safety.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data that may be controlled by third parties, such as SMN topic names, message templates, or notification content.
  • Ingestion points: Resource listings and message data retrieved via hcloud output in scripts/smn_dms_skill.py and displayed to the agent.
  • Boundary markers: None identified in the script's processing of JSON output from the CLI to explicitly distinguish data from instructions.
  • Capability inventory: The skill possesses capabilities to create and delete topics or instances, and publish messages, as defined in SKILL.md and implemented in scripts/smn_dms_skill.py.
  • Sanitization: The script uses list-based command execution which prevents shell injection, but it does not include natural language sanitization to filter instructions embedded within resource metadata or message content.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 06:44 AM
Security Audit — agent-trust-hub — huawei-cloud-smn-dms-message