The Agent Skills Directory

[SAFE]: The skill follows established security protocols for browser-based agents, ensuring that data ingestion from web pages is used only for verification purposes and does not trigger unauthorized actions.
[SAFE]: Credential management is handled through role identifiers rather than plaintext secrets, following security best practices to prevent credential leakage in project files.
[SAFE]: File system operations are scoped strictly to the user-provided project directory for writing reports and artifacts, with an integrity ledger used to maintain the consistency of the verification specifications.
[SAFE]: The skill utilizes a vendor-provided MCP server (amazon-nova-act-mcp) and standard browser interaction tools. No suspicious remote code execution or command injection patterns were identified.
[SAFE]: Multi-layered security checks, including an automated audit pass during compilation, are used to reconcile design specifications with the live application state, preventing the use of orphaned or divergent rules.

ui-verification