ltv-analysis

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

• Third-party content exposure detected (medium risk: 0.30). In Stage 1 Priority 2/3 the runtime builds a JSON “revenue” definition from user-supplied LTV horizon/parameters and then calls analysis adhoc run --model-type revenue --definition '<json>', so the agent ingests indirect outsider-controlled free text/fields that influence the LLM-visible definition.