The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and edit user-supplied prose. Although it lacks explicit boundary markers or sanitization instructions to separate input data from its internal logic, the skill itself possesses no executable capabilities or tool access that could be leveraged for malicious purposes. The primary risk is limited to potential instruction following from malicious data within the text being edited.
[NO_CODE]: The skill package consists entirely of Markdown documents and a YAML configuration file. There are no scripts (e.g., Python, JavaScript, Shell) or binary executables included, meaning no automated operations are performed on the host environment.
[SAFE]: The instructional content focuses on established technical writing principles, such as the Diátaxis framework and Google developer-documentation style. The guidelines emphasize clarity, technical precision, and 'instruction economics'—the efficient use of prompt context to steer agent behavior—without introducing safety risks.

oro