supabase-ts
Pass
Audited by Gen Agent Trust Hub on Mar 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill uses official Supabase libraries and follows industry-standard patterns for authentication, storage management, and database operations. Instructions correctly prioritize security, such as validating JWTs server-side and recommending 'security invoker' for database functions.\n- [DATA_EXFILTRATION]: The skill provides templates for environment variables using placeholders for public keys and includes clear warnings to keep the SUPABASE_SERVICE_ROLE_KEY in server-only contexts, preventing accidental credential exposure.\n- [PROMPT_INJECTION]: The skill facilitates the ingestion of data from external sources (Supabase Database, Storage, and Realtime), which creates a surface for indirect prompt injection.\n
- Ingestion points: Data is fetched via supabase.from().select(), supabase.storage.from().download(), and realtime channel broadcasts (found in SKILL.md, references/storage.md, and references/realtime.md).\n
- Boundary markers: Explicit boundary markers for untrusted data are not provided in the code templates, though structured data patterns are encouraged.\n
- Capability inventory: The skill includes capabilities for CLI operations, database migrations, and Edge Function execution which can interact with external APIs (e.g., OpenAI, Stripe).\n
- Sanitization: The skill strongly advocates for using Zod for schema validation and strict parsing of all external data, which provides a layer of sanitization for ingested content.\n- [COMMAND_EXECUTION]: The skill includes reference guides for the official Supabase CLI for local development and CI/CD workflows, using legitimate and expected commands.
Audit Metadata