The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands via wrapper scripts (convert.sh, convert.ps1, convert.bat) and Python's subprocess module to perform document conversions and environment setup.
Evidence: scripts/convert_document.py uses subprocess.run to call pip install and node scripts.
[REMOTE_CODE_EXECUTION]: The skill automatically downloads and installs external dependencies from official package registries (PyPI and NPM) if they are missing from the user's environment.
Evidence: install_dependencies function in scripts/convert_document.py and _ensure_shared_node_modules which triggers npm install.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it extracts and presents untrusted content from external documents (PDF, DOCX, XLSX, PPTX) directly to the agent's context.
Ingestion points: Untrusted data enters the context through scripts/convert_document.py during the extraction process.
Boundary markers: Absent. The skill does not instruct the agent to use delimiters or ignore instructions found within the converted text.
Capability inventory: The agent has the capability to execute shell commands, write files to the local directory, and access the network via package managers.
Sanitization: The skill performs basic Markdown syntax escaping to ensure formatting consistency but does not filter or sanitize the content for malicious instructions.

docugenius-converter