The Agent Skills Directory

[PROMPT_INJECTION]: Potential for indirect prompt injection exists because the skill processes untrusted requirement documents (PDF, Word, Markdown) to extract test scenarios. Ingestion points: Requirements are read and text is extracted using scripts/extract_document.py. Boundary markers: The skill instructions emphasize structured parsing and the use of Markdown headers and JSON schemas for internal data handling. Capability inventory: The skill utilizes Bash for script execution, Read and Write for file operations, and Glob for discovery. Sanitization: Mitigation is achieved through user-confirmation checkpoints (Phase 2.5, 2.6, 2.8) where the AskUserQuestion tool is used to verify parsed results and resolve ambiguities before any document generation occurs.\n- [COMMAND_EXECUTION]: The skill uses the Bash tool to execute its internal Python scripts (extract_document.py, generate_excel.py, and memory_manager.py) located in the scripts directory to perform document processing and state management.\n- [EXTERNAL_DOWNLOADS]: The documentation provides standard pip install commands for well-known and reputable document processing libraries (openpyxl, PyMuPDF, and python-docx).

generate-test-docs