scientific-critical-thinking
Pass
Audited by Gen Agent Trust Hub on Jun 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill consists of technical guidelines and reference documentation intended for research analysis.
- [PROMPT_INJECTION]: The skill's primary function of analyzing scientific methodology and claims creates a surface for indirect prompt injection. 1. Ingestion points: User-provided scientific papers, abstracts, and experimental design descriptions. 2. Boundary markers: No specific delimiters or "ignore" instructions are mandated for the processed content. 3. Capability inventory: The skill utilizes the Read, Write, and Edit tools. 4. Sanitization: No explicit input validation or filtering of user-provided claims is defined.
Audit Metadata