azure-ai-projects-py
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill adheres to security best practices for AI agent development, emphasizing the use of least-privilege credentials and isolated environments.
- [COMMAND_EXECUTION]: The included evaluation tool (
scripts/run_batch_evaluation.py) employs defensive programming techniques, including the use ofO_NOFOLLOWto prevent symlink attacks and setting restrictive file permissions (0o600) for output files. - [EXTERNAL_DOWNLOADS]: All external dependencies, such as the
azure-ai-projectsandopenailibraries, are official packages from well-known and trusted service providers. - [DATA_EXPOSURE]: The instructions and scripts include explicit safeguards against the accidental logging of sensitive data, such as connection secrets, private traces, or credential chains.
- [INDIRECT_PROMPT_INJECTION]: The skill recognizes the risk of processing external data and implements schema validation and input limits for datasets, while advising developers to treat retrieved documents and tool responses as untrusted data.
Audit Metadata