streamdown
Fail
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: CRITICALINDIRECT_PROMPT_INJECTION
Full Analysis
- Security Configuration Example (Reputation Flag Analysis): Automated scans flagged
assets/examples/custom-security.tsx. Manual review indicates this is a potential false positive; the file provides a template for strict security policies, such as restricting allowed protocols tohttpsandmailtoand limiting domains to a specific allow-list. This is a safe and encouraged pattern for production systems. - Indirect Prompt Injection (Surface Analysis): The skill focuses on rendering content that may originate from untrusted AI sources, which is a known vulnerability surface. The library includes several mechanisms to manage this risk:
- Ingestion point: Untrusted data enters the component via the
childrenprop. - Boundary markers: The library includes a
linkSafetyfeature that requires user confirmation before navigation. - Capability inventory: The system is scoped to markdown rendering and does not perform sensitive file or system operations.
- Sanitization: The implementation uses
rehype-sanitizefor XSS prevention andrehype-hardenfor URL validation by default.
Recommendations
- CRITICAL: 1 file(s) identified as malware by FileRep - DO NOT USE
Audit Metadata