csv-data-summarizer
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Prompt Injection] (LOW): The SKILL.md file employs extremely forceful, adversarial-style instructions (e.g., '⚠️ CRITICAL BEHAVIOR REQUIREMENT ⚠️', 'DO NOT ASK') to override the agent's standard interactive behavior and 'human-in-the-loop' safety patterns.
- [Indirect Prompt Injection] (LOW): The skill is vulnerable to malicious instructions embedded in processed CSV files. 1. Ingestion points: analyze.py (line 15) uses pd.read_csv to ingest untrusted data. 2. Boundary markers: Absent; the data analysis is presented to the agent without delimiters. 3. Capability inventory: The script extracts and returns raw cell contents (e.g., categorical values in analyze.py lines 74-79) which the agent is then instructed to process and report. 4. Sanitization: Absent; malicious strings within CSV cells could potentially trigger unintended agent behaviors during summarization.
Audit Metadata