The Agent Skills Directory

[Indirect Prompt Injection] (MEDIUM): The skill ingests untrusted data from external websites and local project configurations, which could contain malicious instructions designed to manipulate the agent's behavior.
Ingestion points: The skill reads external page content via navigate_page, analyzes performance traces via performance_analyze_insight, and inspects local files like package.json and webpack.config.js in Phase 5.
Boundary markers: There are no delimiters or explicit instructions to treat the ingested data as untrusted, increasing the risk that the agent will follow instructions embedded in the target sites or files.
Capability inventory: The skill has the ability to navigate to arbitrary URLs, perform network analysis, and read sensitive local build configurations.
Sanitization: No sanitization or filtering of external input is specified.
[External Downloads] (MEDIUM): The documentation instructs the user to install a third-party package (chrome-devtools-mcp) using npx. This package is not hosted in a predefined trusted repository or organization, posing a supply chain risk.

web-performance-audit