redline-docx-pl

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to use uvx to download and execute the adeu utility from a remote package registry. This tool is the primary engine for the skill's document processing functionality.
  • [COMMAND_EXECUTION]: The skill executes local Python scripts (memo_negocjacyjne.py, skan_placeholder.py) and the adeu CLI to perform document analysis, redlining, and sanitization.
  • [DATA_EXPOSURE]: The skill includes explicit instructions to use the --author flag when applying changes to documents to prevent the leakage of the local system's account name into document metadata.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect injection as it reads untrusted .docx content. It mitigates this risk by defining a 'Safety Tier' framework that requires verbal or literal human confirmation ('potwierdzam') before performing mutating or destructive file operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 09:44 PM
Security Audit — agent-trust-hub — redline-docx-pl