alibabacloud-openagent-chat

Pass

Audited by Gen Agent Trust Hub on Oct 20, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/bootstrap.py script uses subprocess.run to execute pip install. This is part of a secure bootstrap mechanism that ensures all Python dependencies are installed into a managed virtual environment and verified against the SHA-256 hashes provided in the assets/requirements.txt file.
  • [DYNAMIC_EXECUTION]: The entry point script scripts/openagent.py uses os.execv to switch the process to the Python interpreter within the verified virtual environment. The scripts/package_audit.py development tool also uses compile() to validate Python syntax during the release auditing process.
  • [EXTERNAL_DOWNLOADS]: The skill connects to official Alibaba Cloud API endpoints (e.g., openagent.cn-hangzhou.aliyuncs.com) and downloads official libraries from the standard Python registry during its initial setup. These operations are restricted to vendor-owned and trusted infrastructure.
  • [CREDENTIALS_UNSAFE]: The skill includes defensive instructions and code to manage Alibaba Cloud credentials safely. It explicitly warns against requesting secrets in chat, uses standard SDK credential providers (Environment, CLI Profiles, OIDC, ECS Roles), and provides a credentials inspect command that checks for the presence of keys without ever outputting or resolving the actual secret values.
  • [INDIRECT_PROMPT_INJECTION]: The skill implements a robust validation layer in scripts/openagent_client/validation.py that checks all incoming data from the cloud service against a predefined JSON schema, mitigating the risk of malicious payload injection through manipulated API responses.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 20, 2026, 06:48 AM