The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is designed to ingest and process external text content for conversion. It lacks explicit boundary markers or delimiters to separate the input data from the transformation instructions. While the skill possesses capabilities to write and edit files (Write, Edit, Grep), which could be abused if an input article contained malicious instructions, the highly specific templates and conversion rules provide a structured framework that limits the likelihood of accidental obedience to embedded commands.
[Remote Code Execution] (SAFE): No patterns for downloading or executing remote scripts or binary files were found.
[Data Exfiltration] (SAFE): No network-enabled tools or instructions to access sensitive local system files (e.g., SSH keys, credentials) are present.
[Obfuscation] (SAFE): All content is provided in clear, human-readable Markdown with no evidence of encoded or hidden characters.

zhihu-converter