The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the official aliyun CLI using the subprocess module to perform supplementary queries for DDoS, Cloud Firewall, and billing status. These invocations use argument lists rather than shell strings, and sensitive inputs like Role ARNs are validated with regular expressions in scripts/sts_create.py to prevent command injection.
[CREDENTIALS_SAFE]: Credentials are managed via scripts/sts_create.py, which validates the environment's default credential chain or uses STS AssumeRole. Temporary credentials are cached locally in scripts/.sts_cache.json with restricted file permissions (0o600), and the skill instructions explicitly forbid hardcoding or echoing plaintext keys.
[EXTERNAL_DOWNLOADS]: The skill depends on the official aliyun-python-sdk-core package, which is a standard library provided by the vendor (Alibaba Cloud). No unverified third-party dependencies or remote script executions were detected.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided identifiers (Instance IDs, VSwitch IDs) and passes them to diagnostic scripts. While this creates an attack surface, the risk is mitigated by using these values only as structured arguments for read-only API calls and CLI commands, following a least-privilege model.

alibabacloud-ecs-vpc-publicnetwork-troubleshoot