alibabacloud-emr-starrocks-manage
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs all cloud resource management tasks by executing commands via the official aliyun CLI. It includes specific instructions for setting the User-Agent and switching profiles safely.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface where user-provided inputs are used in shell commands, but mitigates this through comprehensive defensive instructions.\n
- Ingestion points: User input for instance names, IDs, and region identifiers as specified in the SKILL.md and reference templates.\n
- Boundary markers: Explicit validation rules in the SKILL.md defining allowed character sets and naming conventions.\n
- Capability inventory: Management operations like instance creation, scaling, and configuration changes performed through the aliyun CLI across multiple reference files.\n
- Sanitization: Instructions requiring regex validation for all command-line arguments and recommending the use of JSON --body to isolate user input from shell execution.\n- [SAFE]: The skill strictly prohibits downloading or running external scripts, dependencies, or remote URLs. It also defines mandatory masking rules for sensitive fields such as AdminPassword, AccessKeyId, and STS tokens to prevent data exposure in logs or terminal output.
Audit Metadata