The Agent Skills Directory

[SAFE]: The skill's operations are consistent with its stated purpose of performance auditing. It utilizes standard debugging protocols and tools.
[EXTERNAL_DOWNLOADS]: The skill recommends the use of the chrome-devtools-mcp package from the public NPM registry. This is a well-known service, and the package is a standard requirement for the functionality provided.
[COMMAND_EXECUTION]: The instructions include a configuration command (npx -y chrome-devtools-mcp@latest) intended for the user to set up their local environment. This command is presented transparently and requires user action.
[PROMPT_INJECTION]: The skill processes data from external websites via navigation and performance tools. This ingestion of untrusted data is necessary for web auditing and is handled in an analytical, read-only capacity.
Ingestion points: Web page content, network headers, and accessibility trees accessed via navigate_page and performance_start_trace in SKILL.md.
Boundary markers: None identified.
Capability inventory: Browser navigation, performance metric extraction, network request inspection, and accessibility snapshotting.
Sanitization: Not applicable as the skill performs diagnosis of performance metrics rather than executing or interpolating content into sensitive system commands.

web-perf