tooluniverse-immunotherapy-response-prediction
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted patient tumor profiles, including mutation lists and biomarker values, which are interpolated into its analysis phases. \n
- Ingestion points: Patient tumor data is parsed and used throughout the 11-phase workflow defined in SKILL.md.\n
- Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to ignore potentially malicious instructions embedded in the input data during processing.\n
- Capability inventory: The skill uses file-writing capabilities for reporting and Bash/Python execution for scoring and data analysis.\n
- Sanitization: No explicit sanitization or validation of the input biological data is defined before it is processed by the agent's logic or generated scripts.\n- [DYNAMIC_EXECUTION]: The skill explicitly instructs the agent to perform 'COMPUTE, DON'T DESCRIBE' operations, where the agent writes and executes Python code at runtime to process medical statistics and calculate scores. While these scripts follow a predefined analytical template, the dynamic assembly and execution of code based on user-provided data represents a runtime execution surface.\n- [COMMAND_EXECUTION]: The workflow relies on shell access (via Bash) to execute generated Python scripts and facilitate the multi-phase analysis, granting the agent the capability to perform system-level operations to support its clinical computations.
Audit Metadata