syncfusion-angular-inline-ai-assist

Pass

Audited by Gen Agent Trust Hub on Mar 25, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The documentation provides examples that render AI-generated responses directly into the application's DOM using innerHTML or Angular's [innerHTML] property binding. This introduces a surface for indirect prompt injection where malicious HTML or scripts provided by an AI service could be executed in the user's browser context.\n
  • Ingestion points: Untrusted data originates from the AI service response within the onPromptRequest event handler, as demonstrated in SKILL.md and references/events-and-methods.md.\n
  • Boundary markers: The examples do not specify boundary markers or instructions to the agent to ignore embedded instructions in the processed text.\n
  • Capability inventory: The component is designed to modify the UI content of the host application based on AI responses, as shown in the itemSelect and responseTemplate examples in SKILL.md and references/templates-and-toolbars.md.\n
  • Sanitization: The provided code examples demonstrate direct assignment of response strings to DOM properties without explicit sanitization or encoding. Developers using these examples should be advised to implement proper HTML sanitization.\n- [NO_CODE]: The skill consists entirely of documentation and instruction files without including any executable scripts, binaries, or active code components. The risk is limited to the guidance provided for developers.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 25, 2026, 04:41 PM