speckit-tasks

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow explicitly reads outsider-authored free text from specs/<feature>/plan.md and specs/<feature>/spec.md (and optional data-model.md, research.md, contracts/, quickstart.md) via “Load design documents” and then uses their extracted content to generate tasks, so any non-user-authored documents become LLM context.