suggest

Pass

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted user code and natural language queries to identify patterns, creating a potential surface for indirect prompt injection.
  • Ingestion points: Untrusted data is ingested via codeSnippets and userAsk during passive observation.
  • Boundary markers: The instructions lack explicit delimiters or boundary markers to differentiate between the agent's logic and user-provided code.
  • Capability inventory: The skill can invoke the /add command to install external components into the project environment.
  • Sanitization: There is no evidence of filtering or sanitization performed on the ingested code patterns.
  • Mitigation: The skill implements a robust safety control by requiring explicit human consent before triggering any tool-based installations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 17, 2026, 08:24 PM
Security Audit — agent-trust-hub — suggest