The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides several utility scripts for managing textbook infrastructure.
scripts/detect_features.py scans project configuration using standard libraries.
scripts/generate-favicon.py uses the Pillow library to process image files.
scripts/analyze-reading-levels.py evaluates text complexity using the textstat library.
references/kanban-board.md and references/google-analytics.md provide workflows for using official CLI tools like gh and git to manage project metadata.
[EXTERNAL_DOWNLOADS]: The skill references and downloads resources from established and well-known services.
Fetches JS/CSS libraries (KaTeX, MathJax, vis-network, Giscus) from unpkg.com, jsdelivr.net, and giscus.app.
Downloads an author headshot from the author's GitHub Pages domain dmccreary.github.io.
Recommends installation of standard, established Python packages from PyPI (e.g., mkdocs-material, Pillow, textstat).
[PERSISTENCE]: The optional skill-tracker feature modifies the local agent environment for usage analytics.
Installs hooks in ~/.claude/hooks/ and updates ~/.claude/settings.json to log tool usage patterns across sessions, which is its stated purpose.
[DYNAMIC_EXECUTION]: Contains a dynamic module import in the favicon generation utility.
scripts/generate-favicon.py uses __import__ to load the PIL.Image module for image processing, which is a benign use of the pattern for asset creation.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for indirect injection by processing project-specific data files.
Ingests content from docs/course-description.md and learning-graph.json to generate textbook content (glossaries, FAQs, etc.).
While the skill has file-writing capabilities, the data ingestion is part of the intended static site generation workflow.

book-installer