clinical-data-cleaner
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process clinical trial data from external files (CSV, Excel), creating a surface for indirect prompt injection if source data is maliciously crafted.
- Ingestion points: The load_data method in scripts/main.py reads external files specified by the user via the --input CLI argument.
- Boundary markers: The skill does not implement specific prompt delimiters or instructions to ignore embedded commands within the processed data.
- Capability inventory: The skill has capabilities to write files to the local system (cleaned data and audit reports) but lacks network connectivity or arbitrary command execution.
- Sanitization: The processing logic focuses on data type and clinical range validation; it does not sanitize text fields for natural language instructions.
- [EXTERNAL_DOWNLOADS]: The skill references source code and documentation from its vendor repository on GitHub (github.com/aipoch/medical-research-skills).
- [COMMAND_EXECUTION]: The skill instructions provide standard Python CLI commands for executing the data cleaning script and performing pre-execution validation checks.
Audit Metadata