The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to fetch and run the chrome-devtools-mcp package from the npm registry. This package is not provided by a trusted vendor or well-known service.
[REMOTE_CODE_EXECUTION]: By using npx -y package@latest, the skill executes code from a remote source without version pinning or integrity verification, which is a potential vector for supply chain attacks.
[COMMAND_EXECUTION]: The provided evaluate_script tool allows the agent to execute arbitrary JavaScript in the browser context. This capability could be misused if the agent's context is manipulated via external data.
[DATA_EXFILTRATION]: Instructing users to open a remote debugging port (9222) exposes their active browser sessions. A malicious agent or an external process could use this connection to capture sensitive information like session cookies and authenticated site data.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8):
Ingestion points: Untrusted data enters the agent context via navigate_page, list_console_messages, and get_network_request from external websites.
Boundary markers: Absent. There are no instructions to differentiate between user commands and data retrieved from external sources.
Capability inventory: The skill possesses capabilities to write to the browser (evaluate_script, click, fill) and navigate to new origins.
Sanitization: No sanitization is performed on console logs or network data before being processed by the agent.

install-chrome-devtools-mcp