powerpoint

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the pptx_write tool, which accepts a JavaScript module as a string argument. The agent is instructed to write this module at runtime to define the structure and content of the presentation, which is then executed by the tool worker. This mechanism is the core functionality of the skill and is constrained by the PptxGenJS API.
  • [PROMPT_INJECTION]: The skill processes user-provided data and presentation outlines which are subsequently interpolated into the generated JavaScript code, creating an indirect injection surface.
  • Ingestion points: User-supplied presentation content, slide titles, and data passed via the data argument to the pptx_write tool (documented in SKILL.md).
  • Boundary markers: The instructions do not specify the use of delimiters or boundary markers to isolate user-supplied strings within the generated script.
  • Capability inventory: The skill possesses the capability to read and write file system content via the pptx_write, pptx_read, and pptx_template_fill tools.
  • Sanitization: No explicit sanitization or validation steps are outlined for the strings provided by the user before they are embedded into the executable script.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 05:36 AM
Security Audit — agent-trust-hub — powerpoint