genomics-workflow-acceleration
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted genomics workflow files (Nextflow, Snakemake, WDL, Python) to identify optimization targets, creating an indirect prompt injection surface.
- Ingestion points: Workflow paths and repositories provided by the user as described in SKILL.md.
- Boundary markers: No explicit delimiters or boundary markers are defined for user-supplied code blocks.
- Capability inventory: The agent has file modification privileges and shell command execution capabilities.
- Sanitization: The skill includes guardrails against destructive cleanup and clinical variant interpretation, but lacks specific code sanitization for processed inputs.
- [COMMAND_EXECUTION]: The skill uses shell commands to verify system compatibility.
- Evidence: Commands 'nvidia-smi', 'pbrun --version', and 'docker run' are executed for runtime readiness assessment as specified in references/parabricks-runtime-readiness.md.
Audit Metadata