The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/query_dts_tasks.py utilizes subprocess.run() to execute aliyun-cli commands. The parameters, such as region and job type, are validated against a hardcoded whitelist of supported Alibaba Cloud regions, which prevents arbitrary command injection.
[PROMPT_INJECTION]: The SKILL.md file includes forceful instructions using language like 'CRITICAL OUTPUT COMPLETENESS REQUIREMENT' and 'MANDATORY
Model Response Rules'. While these use patterns associated with instruction overrides, they are functionally intended to ensure the AI does not truncate or summarize technical data, preserving the integrity of the reported task lists.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API calls via the aliyun-cli and interpolates it into the agent's context. This creates a potential surface for indirect injection if the source data were untrusted, though the source here is the official vendor API.
Ingestion points: Data enters the system via subprocess.run output in scripts/query_dts_tasks.py which is then parsed as JSON.
Boundary markers: None identified. The skill does not use specific delimiters to wrap the external data within the final markdown output.
Capability inventory: The skill possesses subprocess.run capabilities to invoke the aliyun CLI tool for querying resources.
Sanitization: The script performs data filtering based on the PayType field but does not apply string escaping or content sanitization to fields like DtsJobName before they are rendered into the report.

alibabacloud-dts-task-query