Where Are the Generated SKILL.md Files Stored in book-to-skill?

The book-to-skill pipeline creates a dedicated skill directory for each conversion, storing the master SKILL.md file at ./<skill-name>/SKILL.md by default, with the location fully customizable via the --skill-dir CLI argument.

When converting technical books into AI-ready skill definitions using the virgiliojr94/book-to-skill repository, understanding exactly where the generated SKILL.md files are stored is critical for agent integration. The pipeline automatically constructs a structured output directory during Step 6 of the extraction process, placing the master skill definition alongside supporting chapter chunks and reference materials.

Default Output Location

By default, the tool writes the skill directory alongside the repository root, using the sanitized book title as the folder name. According to the repository's root SKILL.md specification, this occurs during Step 6 of the pipeline—"Create skill directory structure."

The generated directory contains:

• SKILL.md — The master skill definition consumed by AI agents
• chunks/ — Individual chapter-level markdown files (chunk-*.md)
• glossary/, patterns/, cheatsheet/ — Optional supporting artefacts referenced by the master file

For example, converting a PDF titled "MyAwesomeBook" produces:

./MyAwesomeBook/SKILL.md
./MyAwesomeBook/chunks/chunk-01.md
./MyAwesomeBook/glossary/...

Customizing the Output Path

The default location can be overridden using the --skill-dir argument defined in book_to_skill/cli.py. This parameter accepts a custom filesystem path, allowing you to organize generated skills outside the project root or into version-controlled directories.

# Override the default location with a custom directory

python -m book_to_skill.cli /path/to/book.pdf --skill-dir ./generated_skills/MyAwesomeBook

When specified, the pipeline creates the <skill-name> subdirectory within the provided path, maintaining the internal structure described above.

Directory Structure Breakdown

The Master SKILL.md File

Located at /<skill-name>/SKILL.md, this file serves as the primary interface for AI agents. It aggregates metadata, learning objectives, and references to chunked content stored in subdirectories.

Chapter Chunks

The chunks/ subdirectory contains individual markdown files (chunk-*.md) representing granular sections of the source book. AI agents can reference these directly for targeted retrieval or context window management.

Supporting Artefacts

Depending on the book content, the pipeline may generate additional folders:

• glossary/ — Term definitions and domain vocabulary
• patterns/ — Reusable code patterns or architectural decisions
• cheatsheet/ — Quick reference materials

Programmatic Access

To load the generated skill into an AI agent or validation tool, reference the concrete file path using standard library functions. The tools/scan_generated_skill.py utility expects this directory layout when performing safety checks on generated outputs.

# Loading the generated SKILL.md for an AI agent

from pathlib import Path

skill_path = Path("MyAwesomeBook/SKILL.md")
skill_content = skill_path.read_text(encoding="utf-8")

# Pass `skill_content` to your agent's prompt or loader

The tests/test_scan_generated_skill.py suite confirms that the pipeline correctly instantiates this directory structure during execution, ensuring reliable path resolution across environments.

Summary

• Generated SKILL.md files reside in a dedicated skill directory named after the book title, created during Step 6 of the pipeline.
• The default output location is ./<skill-name>/ relative to the repository root, containing the master file plus chunks/ and optional subfolders.
• Override the default path using the --skill-dir argument in book_to_skill/cli.py.
• AI agents consume the master SKILL.md file while optionally referencing granular content in the chunks/ subdirectory.

Frequently Asked Questions

What is the default location for generated SKILL.md files?

By default, the pipeline places the skill directory at ./<skill-name>/ relative to the project root, where <skill-name> is derived from the book title. The master file sits at /<skill-name>/SKILL.md, with chapter chunks stored in /<skill-name>/chunks/.

How do I change where SKILL.md files are saved?

Use the --skill-dir command-line argument when running python -m book_to_skill.cli. This parameter, defined in book_to_skill/cli.py, accepts a custom filesystem path where the pipeline will create the skill directory structure.

What other files are created besides SKILL.md?

The pipeline generates a complete skill package including: chapter-level markdown files in the chunks/ folder; terminology definitions in glossary/; design patterns in patterns/; and quick-reference materials in cheatsheet/. These artefacts are referenced by the master SKILL.md file.

How can AI agents access the generated skill files?

AI agents should target the master SKILL.md file at /<skill-name>/SKILL.md for the complete skill definition. For fine-grained context, agents can directly load individual files from the chunks/ subdirectory, as validated by tools/scan_generated_skill.py.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}