supabase-setup
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Instructions guide the agent to use platform-specific CLI tools like
claude mcp addandcodex mcp addto register the database server. - [INDIRECT_PROMPT_INJECTION]: The skill identifies database records as a source of untrusted data and mandates a 'data-as-data' protocol to prevent instruction execution.
- Ingestion points: Records from the Supabase database.
- Boundary markers: Relies on explicit instruction rather than technical delimiters.
- Capability inventory: Database access and local configuration editing via MCP tools.
- Sanitization: Forbids the agent from interpreting retrieved text as commands.
- [SAFE]: The skill facilitates connection to official Supabase infrastructure, a well-known service provider, using secure OAuth-based authentication.
Audit Metadata