html-throwaway-editor
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data to pre-populate generated HTML editors, creating an attack surface for indirect prompt injection.
- Ingestion points: The skill processes user-provided datasets, configuration files, and transcripts which are then rendered into the generated HTML artifact.
- Boundary markers: The skill includes explicit instructions to treat all returned data as content rather than instructions and requires the use of a
redactSecretsflag during the submission process. - Capability inventory: The generated artifacts are capable of performing local network POST requests via a loopback URL (127.0.0.1) or writing data to the system clipboard.
- Sanitization: The instructions mandate a rigorous secret masking protocol using regex and entropy checks, and strictly prohibit the use of
innerHTMLfor dynamic content to mitigate cross-site scripting (XSS) risks.
Audit Metadata