skills/lidge-jun/cli-jaw-skills/hwp/Gen Agent Trust Hub

hwp

Warn

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The script scripts/hwp_convert.py clones the repository https://github.com/neolord0/hwp2hwpx to a temporary directory and executes its build process using Maven and Java at runtime. This allows code execution from an unverified external source.
  • [COMMAND_EXECUTION]: Several Python scripts run external binaries and system commands using the subprocess module. scripts/hwp_convert.py calls git, mvn, and java. scripts/contact_sheet.py runs pdftoppm. scripts/ooxml/soffice.py executes soffice. scripts/hwpx_cli.py performs process replacement using os.execvp.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface when processing user-provided documents. 1. Ingestion points: scripts/text_extract.py and scripts/hwpx_cli.py extract text from untrusted files. 2. Boundary markers: Absent. 3. Capability inventory: Extensive shell execution capabilities including mvn and git. 4. Sanitization: Document content is not sanitized before presentation to the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 2, 2026, 01:53 PM