bear-hunter-system

Pass

Audited by Gen Agent Trust Hub on Apr 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by processing untrusted user content. 1. Ingestion points: Phase 0 ('What are we learning? Drop your topic, notes, or material'). 2. Boundary markers: Absent from the instructions. 3. Capability inventory: Limited to text analysis and Mermaid diagram generation; no access to subprocesses, file writes, or network operations. 4. Sanitization: No specific input validation or escaping mentioned.
  • [SAFE]: No obfuscation, hardcoded credentials, or persistence mechanisms were detected. The skill is entirely instruction-based and operates within the bounds of text-to-diagram generation.
  • [EXTERNAL_DOWNLOADS]: No remote scripts or external package dependencies identified.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 16, 2026, 04:28 PM