oma-frontend

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

• Third-party content exposure detected (high risk: 0.80). The skill's canonical "Shadcn Workflow" in SKILL.md explicitly instructs the agent to run steps like shadcn_search_items_in_registries and shadcn_get_item_examples_from_registries—actions that fetch and review public registry content (third‑party examples) which the agent is expected to read and use to decide installs/implementations, so untrusted external content can materially influence next actions.