The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided text for stylistic transformation. Ingestion occurs via the agent interface when a user requests a rewrite. While there are no specific boundary markers or sanitization logic defined in the instructions to isolate potential malicious input from the instructions, the skill possesses no capabilities for network access, file system interaction, or command execution across its guidelines, limiting the risk factor strictly to the generated text output.
[EXTERNAL_DOWNLOADS]: The skill metadata lists several external references, including repositories from the author and well-known documentation sources like Wikipedia. These links are provided for informational context and background research for the writing guidelines; they do not involve automated runtime downloads or execution of external code.

deslop-zh