The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the official hcloud CLI to perform all query operations. It includes verification steps like hcloud version and hcloud configure list to ensure the environment is correctly set up.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external data from the Huawei Cloud API, such as WAF policy names, rule descriptions, and configuration details. This data is used to generate plain-language explanations for the user. While the data is controlled by the user's cloud environment, it represents a surface where malicious policy names could theoretically attempt to influence the agent's output.
Ingestion points: hcloud WAF ListPolicy, hcloud WAF ShowPolicy, and various rule-listing commands in SKILL.md.
Boundary markers: The skill uses markdown tables and explicit instruction headers to separate retrieved data from agent instructions.
Capability inventory: The skill uses the hcloud CLI and a file writing tool.
Sanitization: The instructions require the agent to interpret the JSON response into 'plain language' explanations, providing a layer of processing between raw data and output.
[DATA_EXFILTRATION]: The skill provides automated workflows for saving retrieved cloud security configurations to local JSON files. It includes a safety constraint requiring the agent to ask the user for confirmation and path verification before using the platform's native file writing tool. The instructions explicitly forbid using shell redirection (e.g., >) to prevent command injection risks.

huawei-cloud-waf-policy-query