alibabacloud-support
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches ticket processing rules and client updates from the official Alibaba Cloud GitHub repository and official OSS buckets. These downloads target vendor-controlled infrastructure and include integrity checks such as SHA-256 verification.
- [DYNAMIC_EXECUTION]: Implements a self-update mechanism that downloads new versions of the skill's own logic from the vendor's repository. The script performs a syntax check using Python's ast module and verifies version increments before replacing local files.
- [COMMAND_EXECUTION]: Invokes the local Alibaba Cloud CLI via secure subprocess calls to perform ticket operations. It uses a structured command list approach that avoids shell interpolation risks.
- [REMOTE_CODE_EXECUTION]: Contains strings describing the installation of the official Alibaba Cloud CLI using curl and bash. These patterns are provided within a data structure for informational setup guidance and are explicitly marked to not be executed automatically by the agent.
- [INDIRECT_PROMPT_INJECTION]: Ingests customer ticket descriptions and remote rules for processing. The skill mitigates potential injection or data leakage risks by applying a redaction engine to remove secrets and enforcing a strict schema that rejects any rule containing executable commands or scripts.
Audit Metadata