foundry-iq
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- Secure Authentication Integration: The skill integrates with the Azure CLI to obtain ephemeral Entra ID access tokens for service interactions. This approach follows best practices by avoiding the use of persistent secrets, hardcoded credentials, or insecure configuration files.
- Isolated Document Assessment: The skill includes a dedicated assessment worker for processing document samples. This worker uses internal process isolation and Python audit hooks to prevent side effects like network access or unauthorized file modification during parsing. Additionally, it uses the
defusedxmllibrary to protect against XML-based vulnerabilities. - Credential Exposure Mitigations: The ingestion logic contains explicit security controls to prevent the processing of sensitive system files. For example, it checks for and rejects paths associated with SSH keys, AWS credentials, and environment configuration files.
- Verified Resource Provisioning: The skill employs a planning and approval workflow. Infrastructure changes are fingerprinted and require explicit operator confirmation before any state-changing operations are performed, ensuring transparency and reducing the risk of unauthorized resource creation.
- Controlled Subprocess Usage: Python scripts within the skill use subprocesses to interact with the Azure CLI (
az) and for internal worker orchestration. These calls use fixed argument lists and avoid shell execution, minimizing the risk of command injection.
Audit Metadata