The Agent Skills Directory

[SAFE]: The skill provides extensive documentation on security best practices, including row-level security (RLS), parameterized queries to prevent injection, and proper password hashing using Argon2. It also lists known security advisories (GHSAs) for SurrealDB to ensure developers are aware of potential risks in specific versions.
[CREDENTIALS_UNSAFE]: The skill contains hardcoded default credentials (root:root) within a Python test fixture intended for localhost development. However, the skill body explicitly warns against hardcoding credentials in production and provides instructions for secure secret management using environment variables, which is a standard and acceptable practice for educational content.
[INDIRECT_PROMPT_INJECTION]: The skill addresses this risk by providing robust patterns for data validation (using ASSERT and TYPE modifiers) and access control, effectively teaching the AI agent and the user how to build secure interfaces that resist untrusted input.

surrealdb-expert