exploratory-data-analysis
Pass
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes external scientific data files (CSV, FASTA, JSON, etc.) to generate EDA reports, creating a surface for indirect prompt injection if the ingested data contains malicious instructions intended for the agent.
- Ingestion points: The script
scripts/eda_analyzer.pyreads user-provided files via its primary command-line argument. - Boundary markers: Analysis results are structured using a Markdown template and statistical summaries are encapsulated in JSON blocks within the report, providing clear delimiters between data and instructions.
- Capability inventory: The bundled Python script is restricted to local file-system operations (reading inputs and writing reports) and does not possess network or arbitrary command execution capabilities.
- Sanitization: While the script summarizes data and uses structured JSON for reporting, it does not explicitly sanitize the raw content of the data samples included in the final Markdown report.
Audit Metadata