task-manager

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

• Third-party content exposure detected (medium risk: 0.65). In scripts/backlog-groomer.js, the agent calls runMoltbotAssist() with input built from outsider-authored task fields (task.text, task.notes, task.work_notes) and sends them to the local LLM via fetch("http://localhost:5278/api/clawdbot/research", ...), so any free-text in those task fields can become LLM context (indirect prompt injection risk).