The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several shell commands to manage the conversion process, including mkdir, command -v, and the web2md CLI. It also provides instructions for building the tool locally using npm install, npm run build, and npm link.
[EXTERNAL_DOWNLOADS]: The skill is designed to fetch and render content from external URLs using Puppeteer to handle JavaScript-heavy websites. This is the primary intended function of the skill.
[COMMAND_EXECUTION]: The skill interpolates user-provided URLs into shell commands using single quotes (e.g., web2md '<url>'). While single quoting provides a layer of protection against shell injection, URLs containing single quotes could potentially be used to break out of the command string if not further sanitized by the underlying platform.
[INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it ingests untrusted data from external websites and processes it into Markdown for the agent.
Ingestion points: External web content via the web2md tool.
Boundary markers: None explicitly defined in the output instructions to the agent.
Capability inventory: File system access (mkdir, ls, wc), command execution (web2md, npm), and network access (via Puppeteer).
Sanitization: Employs single-quoting for URL inputs in shell commands.

web-to-markdown