azureml-to-zenml-migration
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided AzureML artifacts (code, YAML, environments) to generate a new ZenML project. This data ingestion forms a vulnerability surface where the processing of external content could potentially lead to the generation of unintended or malicious code if the input is adversarial.
- Ingestion points: Phase 1 identifies input from AzureML Python code, YAML components, and asset definitions.
- Boundary markers: The skill does not define strict delimiters for input but instructs the agent to analyze the content before translation.
- Capability inventory: The skill specifies the generation of a full project structure, including executable Python scripts and project configuration files.
- Sanitization: It implements a defensive strategy by flagging unsupported control-flow helpers and complex jobs for human review rather than attempting automatic translation.
Audit Metadata