competitor-analysis

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

• Third-party content exposure detected (high risk: 0.95). The skill’s runtime LLM context is fed by outsider-authored free text from public web pages fetched at runtime via browse cloud fetch in the Gate step (Step 4) and then again during enrichment lanes (Step 5), where fetched markdown/hero text and extracted mentions/benchmarks are incorporated into the agent’s prompts and synthesized outputs.