law-to-markdown
Warn
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script 'scripts/law_to_markdown.py' executes shell commands using 'subprocess.run' to invoke '/usr/bin/osascript' for running JXA/JavaScript conversion scripts.\n- [EXTERNAL_DOWNLOADS]: The skill's documentation and error messages in 'law_to_markdown.py' guide users to download and install an external dependency ('mineru-ocr') from an untrusted GitHub repository ('github.com/cat-xierluo/legal-skills') that is not affiliated with the skill author 'zh-xx'.\n- [DYNAMIC_EXECUTION]: The skill dynamically resolves and executes an external JavaScript file ('convert.js') located in environment-specific directories (e.g., '~/.codex/skills/mineru-ocr/scripts/'). This constitutes dynamic loading from computed paths.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user documents (.txt, .docx, .pdf) and converts them to Markdown, creating a vulnerability surface where malicious instructions embedded in documents could be executed by the agent.\n
- Ingestion points: File paths passed as the 'input' argument to 'scripts/law_to_markdown.py'.\n
- Boundary markers: Absent in the generated Markdown output.\n
- Capability inventory: Shell execution ('subprocess.run'), file system operations ('shutil.copyfile', 'Path.write_text', 'Path.unlink').\n
- Sanitization: Lacks sanitization for prompt injection; focuses solely on structural formatting and whitespace cleanup.
Audit Metadata