The Agent Skills Directory

[SAFE]: The skill employs a local Python script (generate_codebook.py) for all data processing. This deterministic approach ensures that raw dataset content is analyzed on the user's machine without being transmitted to external APIs or the LLM's training context.
[SAFE]: Explicit anti-hallucination guidelines are provided in the instructions, specifically prohibiting the agent from inferring or 'guessing' the meanings of coded variables (e.g., medical grades). It uses a [NEEDS DICTIONARY] flag to mandate human verification from authoritative sources.
[SAFE]: The script uses standard, reputable data science libraries (Pandas, OpenPyXL, PyArrow) for file parsing and does not contain any network-reaching commands (like requests or curl).
[SAFE]: The test suite (test_generate_codebook.sh) follows security best practices by generating synthetic data for regression testing and utilizing secure temporary directory handling.
[SAFE]: No evidence of prompt injection, obfuscation, or unauthorized privilege escalation was found in the skill files or metadata.

generate-codebook