web-perf
Pass
Audited by Gen Agent Trust Hub on Feb 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill suggests configuring the
chrome-devtools-mcptool vianpxfrom the npm registry. This is a standard procedure for enabling the necessary Model Context Protocol (MCP) server for performance auditing. - [PROMPT_INJECTION]: The skill utilizes a workflow that involves ingesting untrusted content from external URLs, creating a surface for indirect prompt injection.
- Ingestion points: Web content and network request data retrieved via the
navigate_pageandlist_network_requeststools. - Boundary markers: No specific delimiters or "ignore instructions" warnings are used to encapsulate output from external websites.
- Capability inventory: The agent is instructed to use browser automation tools and to read local configuration files (such as
package.json,webpack.config.js, andnext.config.js) to identify frameworks and optimization opportunities. - Sanitization: The instructions do not define mechanisms for sanitizing or validating the content retrieved from remote sites before it is processed by the agent.
Audit Metadata