The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to suggest a git clone of a template repository (https://github.com/dmccreary/intelligent-book-template) if a textbook structure is not found. This repository belongs to the skill's author and is used to provide the necessary environment for the graph generation process.
[COMMAND_EXECUTION]: The skill executes several local Python scripts included in the package (analyze-graph.py, add-taxonomy.py, csv-to-json.py, taxonomy-distribution.py, validate-learning-graph.py) to perform data processing, validation, and report generation. These operations are standard for the skill's stated purpose of graph generation.
[REMOTE_CODE_EXECUTION]: The validate-learning-graph.py script checks for the presence of the jsonschema library and prints a recommendation to install it via pip if missing. It does not attempt to automate the installation or execute arbitrary remote code.
[PROMPT_INJECTION]: The instructions contain standard emphasis markers like 'CRITICAL STEP' and 'IMPORTANT' to guide the agent through the multi-step generation process, but no patterns were found that attempt to bypass safety guidelines or override core agent behavior.
[DATA_EXFILTRATION]: No network operations to non-whitelisted or suspicious domains were detected. All file operations (read/write) are confined to the local docs/learning-graph directory created during setup.

learning-graph-generator