video-dashboard

Pass

Audited by Gen Agent Trust Hub on Jul 23, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill defines an explicit 'Untrusted content boundary' to manage data from transcripts and OCR. It instructs the agent to treat all external strings as data only and never as instructions, effectively mitigating indirect prompt injection risks.
  • [SAFE]: The dashboard generation instructions mandate the use of document.createElement() and textContent when rendering external data. This approach prevents cross-site scripting (XSS) by ensuring that untrusted content (like transcript excerpts) cannot be executed as HTML or JavaScript.
  • [EXTERNAL_DOWNLOADS]: The skill installs the chart.js library using npm. It employs security best practices by using the --ignore-scripts flag to block malicious lifecycle scripts and --save-exact to pin the dependency to a specific, reviewed version.
  • [COMMAND_EXECUTION]: A local web server is used for testing via python -m http.server. The server is restricted to the local loopback address (127.0.0.1), which prevents unauthorized remote access to the project files while the dashboard is being viewed.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 23, 2026, 12:47 PM
Security Audit — agent-trust-hub — video-dashboard