skills/iofficeai/aioncore/morph-ppt/Gen Agent Trust Hub

morph-ppt

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill setup process involves downloading and executing installation scripts from https://d.officecli.ai/install.sh and https://d.officecli.ai/install.ps1. These domains are associated with the OfficeCLI tool and its author, iofficeai, and are considered legitimate vendor resources.
  • [COMMAND_EXECUTION]: The skill includes a Python helper script (reference/morph-helpers.py) and various shell scripts (build.sh) that execute the officecli tool. The Python implementation correctly uses list-based arguments in subprocess.run to prevent shell injection vulnerabilities.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from .pptx files, which constitutes an attack surface for indirect prompt injection if the processed files contain adversarial content.
  • Ingestion points: Structured slide data is read in reference/morph-helpers.py using officecli get ... --json calls.
  • Boundary markers: Not present; the tool relies on the binary parsing provided by the underlying CLI tool to isolate content.
  • Capability inventory: The skill has the ability to execute system commands via subprocess.run and modify the filesystem using officecli commands.
  • Sanitization: No specific sanitization of shape names or text content is performed before processing by the helper library, which is consistent with the tool's primary purpose of technical presentation construction.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 08:34 AM
Security Audit — agent-trust-hub — morph-ppt