The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's documentation and troubleshooting section contain instructions for the agent to download and execute shell scripts from an external source using risky patterns such as curl ... | sh and irm ... | iex. Specifically, it points to https://cdn-mineru.openxlab.org.cn/open-api-cli/install.sh and a corresponding .ps1 file. This allows arbitrary code execution from a remote server managed by the Shanghai AI Lab.
[EXTERNAL_DOWNLOADS]: The skill is designed to transmit user-provided documents (PDFs, images, and Office files) to a remote API at mineru.net. While the behavior is disclosed, it involves sending potentially sensitive data to an external third-party service.
[COMMAND_EXECUTION]: The skill functions by executing the mineru-open-api command-line tool with various arguments, including file paths and URLs provided by the user.
[PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection because it processes untrusted documents from local storage or remote URLs. If a processed document contains hidden malicious instructions, the agent might interpret them as valid commands. The skill lacks explicit boundary markers or sanitization logic to mitigate this risk.

pdf-converter