prompt-injection-guard

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 is explicitly designed to gate actions based on content it reads from outsider-authored sources (e.g., fetched web pages/HTML/markdown or scraped docs) and then uses that content to decide whether to run tools, so the LLM context can include that outsider free text via the “summarize or act on HTML/markdown/scraped content” path.