Add Kindle EPUB builder

.gitignore

@@ -0,0 +1,2 @@
+/dist
+/tools/__pycache__

tools/README.md

@@ -0,0 +1,64 @@
+# Tools
+
+This directory contains local build tooling for repository artifacts.
+
+## Kindle EPUB Builder
+
+`build_kindle_epub.py` converts the Markdown wiki into Kindle-compatible EPUB files.
+
+Run it from the repository root:
+
+```bash
+python3 tools/build_kindle_epub.py
+```
+
+The script writes:
+
+- `dist/awesome-time-tracking.epub` - one all-in-one EPUB containing the README and every `details/*.md` page.
+- `dist/split/Awesome-Time-Tracking-Time-Management-Methods.epub`
+- `dist/split/Awesome-Time-Tracking-Personal-Productivity-Tools.epub`
+- `dist/split/Awesome-Time-Tracking-Business-Time-Tracking-Software.epub`
+- `dist/split/Awesome-Time-Tracking-Workforce-Payroll-and-Compliance.epub`
+- `dist/split/Awesome-Time-Tracking-Industry-Specific-Time-Tracking.epub`
+- `dist/split/Awesome-Time-Tracking-Calculators-Templates-and-Utilities.epub`
+
+Each generated book starts with a landing page that includes the repository URL:
+
+```text
+https://github.com/ever-works/awesome-time-tracking
+```
+
+## Split Logic
+
+The split volumes are assigned from `README.md` section headings. For each `##` section, the builder classifies the heading using keyword groups defined in `VOLUMES` inside `build_kindle_epub.py`.
+
+Each linked `details/*.md` page is assigned to the first matching volume where it appears. This avoids duplicate chapters across the split books. Any detail page not linked from the README falls back to the utilities/templates volume.
+
+The beginning-of-book contents page for each split EPUB is grouped by the README sections that contributed articles to that volume. It only links to articles included in that EPUB.
+
+## Markdown Support
+
+The builder intentionally implements a small Markdown subset that matches this wiki:
+
+- headings
+- unordered lists
+- paragraphs
+- fenced code blocks
+- links
+- inline code
+- bold and italic text
+
+Images are converted to their alt text. Internal links to `details/*.md` are rewritten to the generated XHTML files inside the EPUB.
+
+## Validation
+
+Before writing each EPUB, the builder parses every generated XML/XHTML document with Python's XML parser. This catches invalid XML, mismatched tags, and disallowed characters such as null bytes.
+
+After generation, you can also check the EPUB container:
+
+```bash
+unzip -t dist/awesome-time-tracking.epub
+unzip -t dist/split/Awesome-Time-Tracking-Time-Management-Methods.epub
+```
+
+EPUB files can be sent to Kindle using Amazon's Send to Kindle flow.