The Agent Skills Directory

Data Exposure & Exfiltration (LOW): The skill transmits user-provided URLs to the external service mineru.net for processing. While essential for the skill's function, this involves sharing potentially sensitive URLs with a third party.
External Downloads (LOW): The skill's scripts download and extract ZIP files from mineru.net into ~/.openclaw/workspace/mineru/. While the description specifies the contents are Markdown and JSON, the extraction process must be audited for vulnerabilities like Zip Slip (directory traversal via filenames in ZIP archives).
Indirect Prompt Injection (LOW): The skill creates a surface for indirect prompt injection by processing untrusted data from the web.
Ingestion points: External URLs processed via the MinerU API (referenced in scripts/mineru_parse_documents.py).
Boundary markers: The documentation does not specify the use of delimiters or warnings to prevent the LLM from following instructions embedded in the extracted content.
Capability inventory: The skill executes local Python scripts and performs file writes to the user directory.
Sanitization: There is no mention of sanitizing the parsed Markdown before it is returned to the agent.
Command Execution (LOW): The skill relies on running local Python scripts (mineru_parse_documents.py, mineru_extract.py) which interact with the filesystem and network. The full source code for these scripts was not provided for detailed audit.

mineru-extract