web-scraper

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow includes fetching arbitrary public web pages via WebFetch (Phase 2) and then extracting their readable DOM/text (Phase 4), which can contain outsider-authored free text (e.g., malicious prompt-injection content on the target page) that is ingested into the agent’s LLM context.