The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides parameters such as function, evaluate, scripts, and modules that allow the AI agent to execute arbitrary JavaScript code within a remote headless browser context on the Microlink infrastructure.
The function parameter enables the execution of logic with access to the Puppeteer page object and various NPM packages (e.g., @aws-sdk/client-s3, youtube-dl-exec).
The scripts and styles parameters allow for the injection of inline code or absolute URLs into the target page, which could be used for content manipulation or data harvesting.
[PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection (Category 8) because it ingests and processes data from arbitrary external URLs.
Ingestion points: Content is fetched from user-provided URLs in multiple formats including metadata, HTML, and markdown (documented in SKILL.md and api-reference.md).
Boundary markers: The skill does not utilize delimiters or specific instructions to ignore embedded commands within the fetched external data.
Capability inventory: The skill possesses capabilities for remote code execution, asset generation (screenshots/PDFs), and structured data scraping.
Sanitization: No validation or sanitization of the fetched external content is described before the data is returned to the agent's context.
[EXTERNAL_DOWNLOADS]: The skill utilizes the @microlink/mql package and references the @microlink/cli tool. It communicates with external API endpoints at api.microlink.io and pro.microlink.io to perform its functions.
[COMMAND_EXECUTION]: The documentation provides instructions for installing and using the microlink CLI tool, which requires the agent to have permission to execute commands in the local shell environment.

microlink-api