supabase-skill
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the agent interacting with external database data. If the database contains malicious content (e.g., in text fields), it could influence the agent's behavior. The skill includes documentation on mitigating these risks, such as using manual approval for tool calls and employing read-only mode for exploration.
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for setting up MCP servers using
npxto fetch and run the@supabase/mcp-server-supabasepackage from the official npm registry. This is a standard deployment method for the Model Context Protocol and utilizes well-known, trusted organizations. - [DYNAMIC_EXECUTION]: Static analysis flagged documentation in
references/mcp-setup.mdthat discusses SQL injection prevention and sandboxing. The file contains defensive pseudocode examples for input sanitization and operation whitelisting intended to guide secure development, rather than providing exploitable code.
Audit Metadata