antinet-doc-parse
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local files provided via
file_path. - Ingestion points: Reads files in various formats (PDF, Word, Excel) to convert them to Markdown as seen in
SKILL.mdandscripts/run_doc_parse.py. - Boundary markers: None explicitly defined in the instructions to separate parsed content from agent instructions.
- Capability inventory: The skill has capabilities to read local files and write structured JSON output to the file system via
core.runtime.AgentSessionandjson.dumpinscripts/run_doc_parse.py. - Sanitization: No evidence of sanitization or filtering of the parsed text content before it is passed to downstream agent stages.
- [EXTERNAL_DOWNLOADS]: The skill lists external Python dependencies including MinerU, PyMuPDF, pdfplumber, and python-magic. These are standard libraries for document processing and file type identification.
Audit Metadata