manufacturing-industrial

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). The skill can ingest the ToR/RFP and other solicitation/provider text at runtime (outsider-authored free-form procurement solicitation and “discovery” inputs), which would be read into the LLM context via its stated workflow “Load” and required inputs handling.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The workflow explicitly instructs the agent to "Load ../../consulting-frameworks/references/industrial-operations-diagnostics.md", a required runtime-read reference whose content would be injected into the agent context and thus directly control prompts/instructions.