obsidian

Pass

Audited by Gen Agent Trust Hub on Oct 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The documentation provides explicit security guidance for developers, specifically advising against the use of innerHTML and outerHTML to prevent Cross-Site Scripting (XSS) vulnerabilities in favor of safer DOM APIs like textContent.
  • [SAFE]: The skill recommends using Obsidian's requestUrl() API for all network operations to safely bypass CORS restrictions and advises developers to type-check and validate all response data at the API boundary to prevent unsafe data leakage.
  • [SAFE]: Detailed instructions for memory management and component lifecycle handling are provided, promoting the use of registerEvent() and registerDomEvent() to ensure proper cleanup and prevent memory leaks.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes user-supplied plugin code, it includes comprehensive mandatory evidence for safety: it identifies clear ingestion points for untrusted data, provides strong sanitization guidance through its DOM and network API rules, and lacks high-privilege automated execution of processed data, maintaining a safe posture.
  • [SAFE]: The skill defines mandatory accessibility (A11y) standards, including keyboard navigation support, ARIA labels, and focus management, ensuring that developed plugins are inclusive and compliant with industry standards.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 13, 2026, 03:57 PM