procurement-offer-review

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted external data from supplier offers in the OFFER-REVIEW mode. Malicious instructions hidden within these documents could potentially influence the agent's behavior.
  • Ingestion points: SKILL.md describes the agent reading external docx and PDF files provided by suppliers.
  • Boundary markers: The instructions lack explicit boundary markers or warnings to disregard instructions embedded within the processed supplier files.
  • Capability inventory: The agent has the ability to write to the local file system (modifying docx and xlsx files) and execute local Python scripts.
  • Sanitization: There is no mention of sanitizing or validating the input data from supplier offers before processing.
  • [COMMAND_EXECUTION]: The skill uses a custom Python script 'scripts/add_offer_comments.py' to add comments to Word documents. Additionally, the 'references/docx-xlsx-mechanics.md' file describes the use of Windows PowerShell ('powershell.exe') for Excel automation tasks.
  • [EXTERNAL_DOWNLOADS]: The documentation references the installation of standard Python packages such as 'pypdf', 'python-docx', and 'openpyxl' to support its extraction and editing functions. These are well-known and established libraries for document processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 07:09 PM
Security Audit — agent-trust-hub — procurement-offer-review