The Agent Skills Directory

[Prompt Injection] (SAFE): The instructions are limited to linguistic refinement and academic styling. There are no attempts to bypass safety filters, extract system prompts, or override agent constraints.- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file paths, or network communication commands (e.g., curl, wget) were found.- [Indirect Prompt Injection] (LOW): The skill processes untrusted user-provided text. However, it lacks any side-effect capabilities such as file writing, shell execution, or network access, meaning an injection attack would have no functional impact beyond the generated response.- [Remote Code Execution] (SAFE): The skill consists entirely of Markdown and YAML metadata; it does not download, reference, or execute any external packages or scripts.- [Obfuscation] (SAFE): No hidden content, encoded strings, or Unicode-based evasion techniques were detected.

polish