The Agent Skills Directory

[SAFE]: The skill implements document chunking, synthesis, and reporting using only standard Python libraries. All processing occurs locally without network connectivity or external dependencies.
[COMMAND_EXECUTION]: Subprocess calls found in test_chunker.py and test_deepread.py are used exclusively for automated testing of the command-line interfaces for the skill's own scripts.
[DYNAMIC_EXECUTION]: The use of importlib.import_module in test_primitives_present.py is a standard verification step used during testing to ensure local script files are present.
[PROMPT_INJECTION]: The reporting logic in deepread.py includes defensive sanitization via the _escape_cell function, which prevents data extracted from processed documents (such as verbatim quotes) from breaking the markdown table structure in the final output.

deep-read