The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to manage dynamic web content. This includes running a Node.js script for browser automation and using npx to install Model Context Protocol (MCP) tools.
[EXTERNAL_DOWNLOADS]: The skill recommends installing external libraries such as @modelcontextprotocol/server-playwright and playwright. While these originate from well-known sources and are necessary for the skill's functionality, they involve downloading and installing code from external registries.
[REMOTE_CODE_EXECUTION]: The skill uses a fallback mechanism where the agent generates and executes a Node.js script locally to render JavaScript-heavy websites. This constitutes a controlled form of remote code execution during the scraping process.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from external URLs. Malicious instructions embedded within a target website could potentially influence the agent's behavior during the conversion or optimization tasks.
Ingestion points: Content retrieved via WebFetch or the Playwright-based browser rendering script.
Boundary markers: The instructions use Markdown headers and YAML frontmatter to structure the output but lack explicit directives to the agent to disregard instructions contained within the fetched content.
Capability inventory: The skill utilizes file writing (Write), network requests (WebFetch), and shell execution (Bash) capabilities.
Sanitization: No specific sanitization or filtering logic is provided to cleanse the fetched HTML or Markdown content before processing.

web-to-markdown