ai-rag-pipeline

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified in the RAG pipeline patterns.
  • Ingestion points: Data retrieved from web search (Tavily, Exa) and content extraction tools.
  • Boundary markers: Examples do not utilize delimiters (e.g., XML/Markdown blocks) or instructions to ignore commands within external context.
  • Capability inventory: Uses infsh CLI via Bash to execute search and LLM tools.
  • Sanitization: Untrusted web content is directly interpolated into prompt strings without validation or escaping.
  • [COMMAND_EXECUTION]: The skill uses the infsh command within a restricted Bash environment to perform authentication and run AI applications.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 07:32 PM