label-content-name-mismatch
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze external HTML components and markup, which constitutes an untrusted data ingestion point.
- Ingestion points: The instructions specify reviewing "rendered HTML" and "markup" from the user's project.
- Boundary markers: No specific boundary markers or delimiters are provided to protect the agent from embedded instructions in the HTML.
- Capability inventory: The skill itself has no executable scripts or tools, but the agent may have platform-level capabilities like file editing to implement fixes suggested in the instructions.
- Sanitization: There are no instructions for sanitizing the input data.
Audit Metadata