geistdocs

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

• Third-party content exposure detected (medium risk: 0.30). The Geistdocs “Ask AI” feature and its /llms.txt endpoint at runtime expose documentation text to the AI; since those docs/MDX content are derived from tenant-authored repository content (i.e., outsider-authored text can be included in the docs), the agent can ingest free text without selecting a specific trusted item first.