perfect-presentation

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Potential for Indirect Prompt Injection (Category 8) due to the mandatory research phase.
  • Ingestion points: Data enters the context during 'Phase 1 — Deep Research' where the agent is instructed to gather facts, stats, and case studies from external sources.
  • Boundary markers: There are no markers or instructions to treat researched content as untrusted data.
  • Capability inventory: The skill executes shell commands, installs packages, and runs a dynamically generated script (node presentation.js) in Phase 9.
  • Sanitization: No sanitization is performed on the gathered research data before it is interpolated into the JavaScript code that generates the presentation.
  • [COMMAND_EXECUTION]: The skill executes multiple shell commands to facilitate its workflow, including installing dependencies, running the generated presentation script, and performing quality assurance checks using Python modules and system utilities.
  • [EXTERNAL_DOWNLOADS]: The skill triggers the installation of several Node.js packages (pptxgenjs, react, react-dom, react-icons, sharp) from the public npm registry to enable presentation generation and image processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 09:08 PM