The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a feature to visualize cloud architecture by starting a local Python webserver. The implementation uses python3 -m http.server bound to 127.0.0.1, which correctly limits network exposure to the local loopback interface and is standard for development previews.
[INDIRECT_PROMPT_INJECTION]: The skill reads existing configuration and design files from the project workspace during modification tasks. This represents a potential surface for indirect prompt injection if an attacker were to modify these local files. This surface is documented as a risk inherent to infrastructure-as-code management tools.
Ingestion points: .aliyun-ai-ops-spec/{name}/designs/design.md and related .tf files in SKILL.md.
Boundary markers: No specific delimiters are used for the loaded design content.
Capability inventory: Includes cloud resource management via aliyun-cli, file system operations, and shell command execution.
Sanitization: The instructions guide the agent to analyze and summarize the loaded data before continuing the conversation, enabling user oversight.

alibabacloud-planning