nature-paper2word
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the
md-to-word-docxpackage from PyPI if the engine is not found on the system. - [COMMAND_EXECUTION]: The skill operates by executing the
md-to-wordcommand-line utility to perform document conversion, version checking, and path retrieval. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from user-provided academic manuscripts.
- Ingestion points: User-provided content from pasted text, .txt, .md, .docx, or PDF files as described in Section 2 of SKILL.md.
- Boundary markers: The skill defines a specific Markdown 'contract' using style markers (e.g.,
%%), specific heading structures, and YAML configuration blocks to delimit data. - Capability inventory: The skill uses the
md-to-wordtool which performs file reads and writes, and can optionally invoke Word on macOS for PDF proofing. - Sanitization: The skill includes a mandatory 'pre-flight validation' step (Section 4 and 5) that identifies and blocks manuscript content that violates the expected formatting contract.
Audit Metadata