odin

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

• Third-party content exposure detected (high risk: 0.70). The required workflow reads and prints the contents of references/recipes/{slug}.md verbatim into the agent’s LLM context; these recipe files are not shown as user-authored at runtime, so they are treated as outsider-authored free text being ingested via the “read file and print contents” path.