The Agent Skills Directory

[PROMPT_INJECTION]: The skill instructions focus on security best practices and do not contain patterns intended to bypass agent constraints or override safety guidelines.- [DATA_EXFILTRATION]: No evidence of unauthorized network requests or sensitive data harvesting was found. The code examples demonstrate secure handling of tokens using environment variables and standard databases.- [COMMAND_EXECUTION]: The skill does not execute shell commands or use tool-calling patterns that would lead to unauthorized system access.- [REMOTE_CODE_EXECUTION]: No remote script execution or unverified binary downloads are included. All referenced dependencies are standard, well-known software libraries.- [EXTERNAL_DOWNLOADS]: The skill recommends installing standard, legitimate packages for authentication (e.g., jsonwebtoken, fastapi) which are appropriate for its stated purpose.

jwt-implementation