build-review-interface
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of markdown instructions and design specifications. It does not provide any scripts, executables, or code snippets for the agent to run, which significantly limits the technical attack surface.
- [INDIRECT_PROMPT_INJECTION]: The skill describes an interface that ingests and renders untrusted data from LLM traces. It addresses the inherent risk of malicious content (such as XSS or tracking pixels) within those traces by explicitly instructing the developer to sanitize rendered content, strip raw HTML, and disable images in markdown. This demonstrates an awareness of security best practices for handling untrusted data.
- [COMMAND_EXECUTION]: The skill mentions testing the interface with Playwright, a standard automation tool. It does not provide any malicious command strings or unauthorized shell execution patterns.
Audit Metadata