mcp-builder
Warn
Audited by Snyk on Jul 25, 2026
Risk Level: MEDIUM
Full Analysis
MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).
- Third-party content exposure detected (high risk: 0.75). Inside the required runtime evaluation workflow, the harness sends the MCP server’s tool results (which can include arbitrary free text from the target/connected system) back into the LLM message history as
tool_resultcontent (scripts/evaluation.py:116-135), creating an indirect prompt-injection surface.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata