predictive-modeling
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data, creating a potential vector for indirect prompt injection.
- Ingestion points:
scripts/run_fold_table.pyandscripts/leakage_smoke.pyload data via the--inputparameter. - Boundary markers: Absent; the skill does not use protective delimiters when the agent interprets the script outputs.
- Capability inventory:
scripts/run_fold_table.pycan write files to the local system (--out,--predictions-out) and the skill environment allows shell command execution. - Sanitization: Absent; the scripts validate numerical consistency and row integrity but do not sanitize for embedded instructional text in input files.
- [COMMAND_EXECUTION]: The skill instructions require the agent to perform package installations using pip and execute the provided Python scripts in the shell environment.
Audit Metadata