analyze
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes untrusted data from CSV files.
- Ingestion points: All scripts in the scripts/ directory (e.g., descriptive_stats.py, ab_test.py, validate.py) utilize pandas.read_csv() to load data from user-provided file paths.
- Boundary markers: Absent. The instructions in SKILL.md do not define explicit delimiters or 'ignore' directives to isolate data content from the agent's instruction logic.
- Capability inventory: The skill uses plt.savefig() in scripts/chart_templates.py to write files to the system and is configured with Bash tool permissions in SKILL.md to execute the Python analysis scripts.
- Sanitization: Partial. While validate.py checks for data anomalies and logical pitfalls, it does not perform sanitization of text fields to filter out potential embedded prompt injection instructions.
Audit Metadata