The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell commands to locate its script bundle and to run playwright-cli for browser automation. These operations are restricted to the local environment and are necessary for the skill's page analysis functionality.- [DYNAMIC_EXECUTION]: The skill utilizes playwright-cli eval to execute analysis logic within the browser context. The code being executed is a local JavaScript bundle provided with the skill, and the process is managed through a temporary configuration file.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted web content as its primary function. It includes a specific warning to the user about external content and implements protective measures such as truncating extracted text to 30 characters and using unique positional IDs for element mapping.

page-tree