thesis-defense-pptx

Pass

Audited by Gen Agent Trust Hub on May 18, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on several PowerShell scripts (e.g., clone_template_deck.ps1, export_pptx_png.ps1) that use COM automation to control Microsoft PowerPoint. These scripts are invoked with the -ExecutionPolicy Bypass flag, which is a standard procedure for local task automation within this context.
  • [PROMPT_INJECTION]: The skill processes external, user-provided documents (PDF and LaTeX), creating a surface for indirect prompt injection. Ingestion points: scripts/extract_thesis_context.py extracts text and figures from thesis projects. Boundary markers: No specific delimiters or instructions are used to separate untrusted document content from system prompts. Capability inventory: The agent has the ability to execute shell commands, perform file system operations, and manipulate local applications. Sanitization: Content undergoes basic text cleaning but no semantic filtering or safety-specific validation.
Audit Metadata
Risk Level
SAFE
Analyzed
May 18, 2026, 02:50 PM
Security Audit — agent-trust-hub — thesis-defense-pptx