canva

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

• Third-party content exposure detected (high risk: 0.80). The skill calls the public Canva Connect API (e.g., scripts/canva.sh and SKILL.md use https://api.canva.com/rest/v1 endpoints such as GET /designs and GET /designs/{id}), which returns user-generated designs/assets from api.canva.com that the agent fetches and reads as part of its workflow.