The Agent Skills Directory

[DATA_EXFILTRATION]: The skill is designed to send local file content and external URLs to a third-party service (markdown.new) for conversion. While this is the primary function of the tool, it inherently involves sending data to an external server, which poses a privacy risk if used with sensitive local files or confidential URLs. The author has included explicit warnings in the documentation and script output to mitigate accidental misuse.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from arbitrary external websites and search results, which is a common vector for indirect prompt injection.
Ingestion points: The scripts/markdown_convert.py script fetches data from URLs provided to the convert, crawl, and search commands.
Boundary markers: The skill does not implement technical boundary markers or 'ignore' instructions in the output, though it provides prominent warnings to the user and agent in SKILL.md and via stderr in the Python script.
Capability inventory: The skill can read local files (convert_file), perform network requests (urllib.request), and write results to the local filesystem if an output path is provided.
Sanitization: The script does not sanitize or filter the Markdown content returned by the external service; it directly passes the raw or JSON-extracted content to stdout or a file.

markdown-new