The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions specify running an external MCP server and CLI tools using the npx kordoc and npx -y kordoc mcp commands. This involves downloading and executing code from the npm registry originating from a third-party repository at github.com/chrisryugj/kordoc.
[COMMAND_EXECUTION]: The skill documentation provides several shell command examples for parsing documents and filling forms (e.g., npx kordoc 보고서.hwp -o 보고서.md), which allows for execution in the host environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted document files (HWP, HWPX, PDF, XLSX, DOCX) as its primary function.
Ingestion points: Untrusted data enters the agent context through tools like parse_document, parse_pages, and parse_table mentioned in SKILL.md.
Boundary markers: The instructions do not define delimiters or specific warnings for the agent to ignore instructions embedded within the extracted text.
Capability inventory: The skill can execute commands via npx and perform file system operations.
Sanitization: There is no description of sanitizing or filtering the parsed content before the agent processes it.

doc-reader