document-parser

Parse and extract content from .docx, .pdf, and .txt documents. Extracts plain text and tables for analysis. Use when the user uploads a document file or ask...

This skill appears to implement a legitimate document parser, but packaging issues make it risky to install and run blindly. Before using: (1) inspect main.py — it currently contains a syntax error (the if __name__ guard is malformed) and differs from parse_document.py; (2) prefer running scripts/parse_document.py (the README and SKILL.md point to it) in a sandboxed environment to confirm behavior; (3) run the install script in a virtualenv rather than system Python; (4) test with non-sensitive documents first to ensure output and that no unexpected network activity occurs; (5) if you plan to let the agent invoke the skill automatically, consider fixing or removing the broken main.py and correcting skill.yaml to point to the intended entry to avoid accidental execution of a different/broken module. If you need higher assurance, ask the publisher for a corrected package or source provenance (homepage/author contact).

Type: OpenClaw Skill Name: mjk39966-document-parser Version: 1.0.0 The document_parser skill bundle is a legitimate tool for extracting text and tables from .docx, .pdf, and .txt files. It uses standard Python libraries (python-docx, PyPDF2, pdfplumber) and contains no evidence of data exfiltration, malicious execution, or prompt injection. A minor syntax error exists in main.py (missing '==' in the __main__ check), but this is a functional bug rather than a security risk.