vertexai-to-zenml-migration
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data, specifically Kubeflow Pipelines (KFP) DSL source files, Vertex submission code, and compiled YAML/JSON pipeline templates. This data is used to drive the generation of new Python code and execute project initialization commands.
- Ingestion points: Receives KFP DSL source, Vertex submission code, and compiled pipeline templates (YAML/JSON) in 'SKILL.md' (Phase 1).
- Boundary markers: The instructions lack specific boundary markers or 'ignore embedded instructions' delimiters when the LLM processes these user-provided artifacts.
- Capability inventory: The skill instructs the agent to write multiple Python files ('steps/.py', 'pipelines/.py', 'run.py') and execute shell commands ('zenml init') at the project root.
- Sanitization: There are no instructions to sanitize or validate the content of the external artifacts before they are used to influence the generated code structure or logic.
- [COMMAND_EXECUTION]: The skill directs the agent to execute shell commands locally to initialize the project environment.
- Evidence: Instructions in 'SKILL.md' (Phase 3) mandate running 'zenml init' at the project root as part of the migration workflow.
Audit Metadata