alibabacloud-iac-code

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads a runtime manifest and binary packages from Alibaba Cloud's official OSS domain (ros-public-tools.oss-cn-beijing.aliyuncs.com). This is normal vendor functionality for the 'aliyun' author.
  • [REMOTE_CODE_EXECUTION]: The skill executes a downloaded binary (iac-code). This is performed securely through a multi-stage verification process: the manifest digest is hardcoded and verified using SHA256, and the binary artifact's digest is verified against the manifest before execution.
  • [DYNAMIC_EXECUTION]: The script uses subprocess.Popen and subprocess.run to manage the local A2A server and perform version checks. These calls use list-based arguments and validated paths, avoiding shell injection risks.
  • [DATA_EXFILTRATION]: The bridge script includes robust sanitization logic (_SECRET_PATTERN, _SENSITIVE_KEY_PATTERN) to redact tokens, passwords, and API keys from logs and agent-visible output.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes user prompts via files, it enforces strict protocol boundaries and requires explicit user approval for permissions, candidate selection, and deployment actions, significantly mitigating injection risks.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 05:42 AM
Security Audit — agent-trust-hub — alibabacloud-iac-code