The Agent Skills Directory

[COMMAND_EXECUTION]: The run.js script functions as a universal executor that takes JavaScript code (either as an inline string or written to a file), saves it to a temporary file named .temp-execution-[timestamp].js, and executes it using the Node.js require() function.
[COMMAND_EXECUTION]: The detectDevServers function in lib/helpers.js uses the Node.js http module to programmatically probe a list of common ports (e.g., 3000, 3001, 8080) on localhost to identify running services.
[EXTERNAL_DOWNLOADS]: The setup script in package.json and the installPlaywright function in run.js invoke npm install and npx playwright install to download the Playwright framework and Chromium browser binaries from public package registries.
[INDIRECT_PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection as it is designed to browse and extract data from external websites.
Ingestion points: Data enters the agent context through browser navigation in SKILL.md and content extraction functions like extractTexts and extractTableData in lib/helpers.js.
Boundary markers: There are no explicit delimiters or instructions provided within the skill's scripts to treat the extracted web content as untrusted or to ignore embedded instructions.
Capability inventory: The skill possesses high local capabilities, including the ability to execute arbitrary JavaScript code via run.js and perform network requests.
Sanitization: The skill does not implement sanitization or filtering on the text extracted from web pages before it is potentially processed by the agent.

playwright-skill