Supabase Developer

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflows explicitly read and act on user-generated, potentially untrusted content—e.g., selecting and subscribing to posts/comments (supabase.from('posts').select(...), real-time subscriptions), downloading/displaying files from storage buckets (storage.from(...).download / getPublicUrl), and performing semantic search/embeddings over post content—which the agent would ingest and use to drive actions and responses.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The Supabase Edge Function example imports remote modules at runtime (which are fetched and executed) — specifically https://deno.land/std@0.168.0/http/server.ts and https://esm.sh/@supabase/supabase-js@2 — and the function depends on those external modules to run.