electric-proxy-auth
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions promote security-first development by providing explicit warnings against common pitfalls and showing correct implementation patterns for Electric SQL proxies.
- [INDIRECT_PROMPT_INJECTION]: The skill demonstrates how to handle external request data safely by implementing a whitelist-based parameter forwarding strategy. It identifies the risk of clients controlling database query parameters and provides code to filter inputs against a known safe set of protocol parameters (
ELECTRIC_PROTOCOL_QUERY_PARAMS). - Ingestion points:
request.urlinSKILL.md. - Boundary markers: The implementation uses a explicit whitelist to restrict which client-provided parameters are forwarded.
- Capability inventory: The skill uses
fetch()to forward requests to a server-sideELECTRIC_URL. - Sanitization: Input parameters are strictly filtered against the
ELECTRIC_PROTOCOL_QUERY_PARAMSarray. - [CREDENTIALS_SAFE]: The skill provides critical guidance on secret management, warning against exposing
ELECTRIC_SECRETorSOURCE_SECRETto the browser and demonstrating how to securely inject them using server-side environment variables. - [COMMAND_EXECUTION]: The skill includes documentation on performance tuning for the Bun runtime, specifically how to increase
BUN_CONFIG_MAX_HTTP_REQUESTS. This is a standard operational command and does not introduce security vulnerabilities.
Audit Metadata