The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses instructional language consistent with technical documentation. While it uses the word 'CRITICAL' to emphasize language-specific features (1-based array indexing), it does not attempt to override safety guidelines or bypass system constraints.
[DATA_EXPOSURE_AND_EXFILTRATION]: No sensitive file paths, hardcoded credentials, or unauthorized network operations were found. The code examples demonstrate secure practices such as parameter binding in SQL queries.
[OBFUSCATION]: All content is in plain text. No Base64, hex encoding, zero-width characters, or homoglyph attacks were detected. The included URLs are clear and point to the official documentation for the technology.
[EXTERNAL_DOWNLOADS]: The skill does not contain instructions to download or execute external scripts, packages, or binaries.
[DYNAMIC_EXECUTION]: The documentation does not demonstrate or encourage unsafe dynamic execution patterns or the use of untrusted inputs in sensitive functions.

boxlang-language-fundamentals