kordoc-korean-document-parser
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install multiple Node.js packages, including
kordoc,kordoc-mcp,pdfjs-dist,tesseract.js, andglob, usingnpm installandnpx. - [COMMAND_EXECUTION]: The skill demonstrates the execution of shell commands through
npxfor both CLI document conversion and the setup of an MCP server. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a parser for untrusted HWP, HWPX, and PDF documents. If the resulting Markdown output is used as input for an agent's subsequent reasoning steps, it could introduce malicious instructions embedded in the original document.
- Ingestion points: Processing of local document files via
readFileSync,glob, or the programmaticwatchAPI. - Boundary markers: The provided code examples do not specify the use of delimiters or specific instructions to the agent to ignore content within the parsed Markdown.
- Capability inventory: The skill has the capability to write to the local file system (
writeFileSync) and trigger network requests (via thewebhookparameter in thewatchfunction and CLI). - Sanitization: There is no mention of sanitizing or escaping the content extracted from the documents before it is output as Markdown.
Audit Metadata