kordoc-parse-document
Warn
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses
npm exec --package=kordocto download and install third-party packages from the public NPM registry at runtime. - [REMOTE_CODE_EXECUTION]: The skill executes the downloaded
kordocpackage via shell command execution. Whilepdfjs-distis a well-known service,kordocis an unverified third-party utility, posing a risk if the package were to be compromised or malicious. - [COMMAND_EXECUTION]: The skill executes shell commands that incorporate user-provided file paths. While instructions suggest resolving to absolute paths, this remains a significant capability.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to parse and summarize content from external HWP, HWPX, and PDF documents, which could contain malicious instructions.
- Ingestion points: Document content parsed from local files in
SKILL.mdworkflow. - Boundary markers: None identified; the skill summarizes parsed content directly into the agent context.
- Capability inventory: File system access and shell command execution via
npm exec. - Sanitization: No evidence of sanitization or escaping of the parsed document content before it is processed by the agent.
Audit Metadata