Cross-Reference Check

Validate all cross-references across the repository.

Step 1: Run the Automated Checks

Execute the validation script:

python3 .claude/skills/cross-reference-check/scripts/validate.py

The script checks:

• Internal .md link resolution (do all [text](path.md) links point to existing files?)
• Chapter H1 title format (does every chapter start with # Chapter?)
• Chapter listing in README.md (is every chapter file referenced?)
• Template path consistency (do paths referenced in guide chapters exist?)
• Checklist file references (is every checklists/*.md file referenced somewhere?)
• Canonical term presence (do CLAUDE.md and README.md contain required terms?)
• Chapter title wording (do README.md and CHEATSHEET.md entries match each chapter's H1?)
• Next: footer title match (does each Next: link's text match the target chapter's H1?)
• Canonical terms registry drift (does the script's dict match references/canonical-terms.md?)

Output is pipe-delimited: TYPE|FILE|LINE|ISSUE|FIX. If all checks pass, the script exits 0.

False positives are expected. The script cannot distinguish real links from illustrative examples (e.g., `[Chapter 02](02-project-bootstrap.md)` shown as a formatting convention, or [text](path.md) used as a pattern example). Review each finding before reporting it — read the surrounding context to determine whether the flagged link is an actual reference or example text.

Step 2: Report

Present all findings (automated + manual) as a table:

File	Line	Issue	Fix
...	...	...	...

If everything passes, output: All cross-references valid.