The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to configure a tool that fetches the chrome-devtools-mcp package from the NPM registry, an official and well-known package repository.
[REMOTE_CODE_EXECUTION]: The skill recommends using npx -y chrome-devtools-mcp@latest, which involves downloading and executing third-party code from the NPM registry to provide the necessary auditing capabilities.
[PROMPT_INJECTION]: The skill ingests untrusted data from audited websites via tools like navigate_page and take_snapshot, creating a surface for indirect prompt injection.
Ingestion points: The agent processes HTML, network traces, and accessibility trees from target URLs defined at runtime.
Boundary markers: No delimiters or specific instructions to ignore embedded commands within the audited content are present.
Capability inventory: The skill utilizes navigate_page, performance_start_trace, list_network_requests, and take_snapshot to extract data from external sources.
Sanitization: There are no instructions for validating or sanitizing the content retrieved from external websites before processing.

web-perf