interactive-reveal-deck
Warn
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: MEDIUMPRIVILEGE_ESCALATIONDYNAMIC_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The skill instructions and documentation advise executing PowerShell scripts using the -ExecutionPolicy Bypass flag, which is a technique used to circumvent local security restrictions on script execution.\n- [DYNAMIC_EXECUTION]: The assets/js/widgets.js file utilizes the new Function constructor to evaluate mathematical and logic expressions stored in data-calc and data-bar HTML attributes. This facilitates the execution of JavaScript code derived from the presentation content.\n- [COMMAND_EXECUTION]: The scaffolding and testing scripts perform various command-line operations, including npm install for dependency management, Start-Process to launch a local web server (http-server), and executing agent-browser for automated slide checks.\n- [INDIRECT_PROMPT_INJECTION]: The skill has an indirect injection attack surface as it processes slide data while possessing powerful execution capabilities.\n
- Ingestion points: Slide content and custom data attributes in assets/index.html are ingested and processed by the JavaScript widget framework.\n
- Boundary markers: No explicit markers or instructions are present to distinguish between data and potential malicious commands in the slide source.\n
- Capability inventory: The skill performs file system modifications, subprocess execution, and network operations (via npm) within scripts/New-RevealDeck.ps1 and scripts/Test-DeckLayout.ps1.\n
- Sanitization: No evidence of input validation or sanitization was found for the expressions passed to the dynamic evaluation functions.
Audit Metadata