therapy-session

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

• Third-party content exposure detected (medium risk: 0.65). The workflow runs Grep/Read within a user-chosen folder or contributor’s files, which can include outsider-authored free text such as commit messages, README/docs, or other repository text written by other people; that text is ingested into the agent’s LLM context during the “Quick scan” step.