funnel-analysis
Pass
Audited by Gen Agent Trust Hub on Apr 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and metadata contain no prompt injection, obfuscation, or malicious intent. All operations are focused on the primary purpose of conversion funnel analysis.- [SAFE]: The included Python snippets use trusted libraries (pandas, numpy, matplotlib, seaborn) to perform local data analysis on user-provided CSV files. No remote code execution or exfiltration patterns were detected.- [SAFE]: Indirect Prompt Injection Surface: The skill ingests untrusted data from 'user_events.csv'. 1. Ingestion points: user_events.csv (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: Data visualization and metrics calculation. 4. Sanitization: Absent. The risk is considered minimal given the structured nature of the data processing.
Audit Metadata