design-review

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md instructs the skill to “Read the target design document in full” and to pass that GDD text (plus related referenced docs and prior review logs) into spawned subagents’ prompts, so any outsider-authored content in those documents can become LLM-readable free text in the agent context.