The Agent Skills Directory

[SAFE]: The skill provides exhaustive guidance on building Manifest V3 extensions, adhering to official Google Chrome developer best practices. It covers critical security aspects such as correctly handling service worker persistence and message passing channels.- [COMMAND_EXECUTION]: The skill includes benign utility scripts for developers, such as a Python/Node.js icon generator and a Bash packaging script. These are standard development tools intended to be run by the developer during the build process.- [REMOTE_CODE_EXECUTION]: The documentation provides secure patterns for executing dynamic or user-provided code using the extension sandbox and the userScripts API. It correctly identifies CSP restrictions and instructs the AI to use secure communication (postMessage) and explicit user opt-in (Developer Mode) as required by the platform.- [CREDENTIALS_UNSAFE]: The skill explicitly warns against hardcoding API keys and sensitive credentials in published extensions. Examples use clearly marked placeholders (e.g., 'YOUR_API_KEY_HERE') or public identifiers (OAuth client IDs) as expected in development guides.

chrome-extensions