task-workflow-new

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 ingests repo-local task inputs from .tasks/task.md, .tasks/domain.md, and especially every file under .tasks/files/ (Phase 1 “Task Intake And Codebase Research”), which are outsider-authored free text in the general case because the operating user did not author them; this text is then used to generate/drive the agent’s LLM context for planning and subsequent phases.