The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes data from external websites, which serves as a potential vector for indirect prompt injection if those sites contain malicious instructions hidden in metadata or comments.
Ingestion points: The skill ingests data through browser automation tools defined in references/devtools.md, including navigate_page, take_snapshot, and performance_start_trace.
Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions that might be embedded in the captured trace or snapshot data.
Capability inventory: The skill has the capability to navigate to arbitrary URLs, capture screenshots, and list network requests.
Sanitization: There is no evidence of sanitization or filtering applied to the content retrieved from target websites.
[EXTERNAL_DOWNLOADS]: The skill references external documentation from a well-known service to guide its performance audits.
Evidence: The file references/devtools.md contains a link to the official Chrome DevTools MCP repository on GitHub (github.com/ChromeDevTools/chrome-devtools-mcp) for tool reference documentation.

web-perf