The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the aliyun CLI to interact with Alibaba Cloud WAF and SLS services. These calls are implemented securely in scripts/get_waf_logs.py using argument lists with subprocess.run, which prevents command injection.
[DATA_EXFILTRATION]: The skill retrieves log entries but includes robust logic in the Python script to mask sensitive fields, including IP addresses, User-Agents, and URI query parameters, ensuring that PII is protected in the analysis report. No exfiltration to external third-party domains is performed.
[CREDENTIALS_UNSAFE]: No hardcoded credentials are used. The skill is designed to use the default Alibaba Cloud authentication chain, such as local configuration files or RAM roles.
[PROMPT_INJECTION]: The skill processes user-provided inputs like the Request ID. It implements strong validation via regex (^[a-zA-Z0-9-]{1,128}$) in scripts/get_waf_logs.py and emphasizes manual confirmation for rule status changes.
Ingestion points: User-provided Request ID and automated SLS log retrieval.
Boundary markers: Delimiters are not explicitly used in prompts, but input is strictly typed and validated.
Capability inventory: Execution of aliyun CLI for log queries and WAF rule status modification.
Sanitization: Use of regex-based allow-lists for all dynamic input parameters.

alibabacloud-waf-checkresponse-intercept-query