The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes external web pages, creating a vulnerability surface where malicious instructions embedded in third-party content could influence agent behavior.
Ingestion points: The skill ingests untrusted data from the DOM, network headers, and performance traces of target URLs using Chrome DevTools MCP tools.
Boundary markers: No explicit delimiters or instructions are provided to the agent to distinguish between the skill's instructions and the content being analyzed.
Capability inventory: The skill possesses powerful capabilities including execute_command, write_file, and wrangler_deploy as defined in the allowed-tools metadata.
Sanitization: Although the 'Security Guardrails' section (ASI05) advises against constructing scripts from external content without sanitization, the prompt instructions do not provide specific mechanisms to implement this enforcement.
[EXTERNAL_DOWNLOADS]: The skill references and retrieves configuration data and performance thresholds from external documentation sites.
Fetches technical guidelines and metric definitions from Google's official web.dev and Chrome DevTools documentation repositories.

web-perf