office-remediation
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides patterns for PowerShell COM automation (e.g.,
New-Object -ComObject Word.Application) and shell commands (sed,zip,unzip) to programmatically modify Office documents. These are standard techniques for the skill's stated purpose. - [INDIRECT_PROMPT_INJECTION]: The skill provides patterns for processing external Office documents (Word, Excel, PowerPoint). Ingesting untrusted document content poses a potential surface for indirect prompt injection if the agent interprets document text or metadata as instructions.
- Ingestion points: File loading calls in
SKILL.mdviapython-docx,openpyxl, andpython-pptxlibraries. - Boundary markers: Absent; the provided code snippets do not include delimiters or warnings to ignore instructions inside the documents.
- Capability inventory: The skill documents file system writes (
.save()), PowerShell COM automation, and direct XML manipulation viasedacross Word, Excel, and PowerPoint files. - Sanitization: Absent; the patterns focus on functionality and do not demonstrate input validation or sanitization of document content.
Audit Metadata