The Agent Skills Directory

[SAFE]: The skill consists of documentation and schemas for remote API calls to a vendor-hosted service. No executable code or scripts are included in the skill package itself.
[PROMPT_INJECTION]: The skill introduces a surface for indirect prompt injection because it is designed to ingest and convert untrusted external data (such as PDFs, HTML, or uploaded files) into text that the agent processes.
Ingestion points: SKILL.md defines the convert action, which accepts external data via the url, file_id, and file_base64 parameters.
Boundary markers: Absent. The instructions do not mandate the use of delimiters or 'ignore' instructions when the agent handles the converted Markdown output.
Capability inventory: The agent typically uses the output for analysis or as input for further actions, creating a potential path for embedded instructions in files to influence subsequent logic.
Sanitization: Not present. The skill provides direct conversion of source content without explicit sanitization or filtering instructions for the agent context.

markitdown-hosted-markdown-generator