ai-chat

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow renders chat message content supplied by the user and/or assistant into the UI via Angular’s StreamingMessageComponent ([innerHTML]="renderedContent()"), where renderedContent is produced by MarkdownService.render(...) (using marked.parse) and then trusted with DomSanitizer.bypassSecurityTrustHtml, so outsider-authored free text can reach LLM-ingestion/HTML-rendering without selecting a specific item first.