The Agent Skills Directory

[SAFE]: The skill documents the implementation of authentication handlers, gateway configuration, and endpoint protection within the Encore.ts framework.
[CREDENTIALS_UNSAFE]: No hardcoded credentials or secrets were detected. The skill correctly demonstrates the use of the encore.dev/config.secret utility for managing sensitive information like JWT secrets.
[EXTERNAL_DOWNLOADS]: Dependencies listed in the code examples, such as jose for JWT validation and vitest for testing, are standard, well-known libraries in the Node.js ecosystem. The internal imports (e.g., encore.dev/* and ~encore/*) are consistent with official framework practices.
[COMMAND_EXECUTION]: Database queries shown in the examples use tagged template literals (e.g., ``db.queryRow`... WHERE api_key = ${apiKey}```), which is the standard method in Encore.ts for creating prepared statements to prevent SQL injection.
[INDIRECT_PROMPT_INJECTION]: While the skill describes handling external inputs like Authorization headers and cookies, it incorporates validation steps (JWT verification, session lookup) and does not interpolate untrusted data into executable commands or LLM prompts.
[OBFUSCATION]: No hidden characters, encoded strings, or other obfuscation techniques were found.

encore-auth