The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/robust_loader.py utilizes importlib.import_module to dynamically check for and load optional dependencies such as openpyxl, xlrd, pandas, pypdf, and pdfplumber. This is implemented as a safety feature to provide informative error messages if required libraries are missing from the execution environment.
[COMMAND_EXECUTION]: The orchestration script scripts/run_pipeline.py uses subprocess.run to execute other internal scripts (robust_loader.py, clean_data.py, etc.) in a sequential pipeline. The commands are constructed using sys.executable and localized script paths, representing standard process management for a multi-step workflow.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data files from problem_files/ and crawled_data/ directories. While this presents a surface for indirect prompt injection if data contains malicious instructions, the risk is low as the data is parsed as structured content using pandas and json libraries rather than being executed as instructions.
[SAFE]: The skill adheres to best practices by scoping file operations to a dedicated paper_output/ directory and utilizing a manifest-based approach (input_manifest.json) to validate inputs before processing.

data-cleaning-and-visualization