feat: initialize wiki-referencify prototype CLI tool

[konard]

Summary

Implements the prototype of the globally installable npm CLI tool wiki-referencify that covers most concepts/terms in a markdown document with Wikipedia links.

Fixes #1

What was implemented

Core algorithm

• Longest-match first: Prefers finding the longest sequences of words that match a single Wikipedia article (e.g., prefers "Machine learning" over "Machine" + "learning" separately)
• Greedy approach: Processes tokens left-to-right, taking the longest matching phrase at each position
• Disambiguation page support: If a matched term has a disambiguation page (e.g., Atlas_(disambiguation)), links to the disambiguation page first — per spec
• Auto-disambiguation mode: With --auto-disambiguation flag, collects contexts from exactly matched articles and skips disambiguation pages for a cleaner output

Wikipedia API integration

• Uses the real Wikipedia API (https://en.wikipedia.org/w/api.php) to check if phrases exist as articles
• File-based caching in system temp dir to avoid repeated API calls
• Batches up to 50 titles per API request
• Handles normalization and redirects correctly

Markdown-aware processing

• Skips headers (ATX # style and setext ===/--- style)
• Skips existing links (doesn't re-link already linked text)
• Skips code blocks (fenced ```, inline `, and indented)
• Skips table headers (first row of markdown tables)
• Respects formatting markers (bold **, italic * — prevents phrases from spanning across formatting boundaries)

Filtering

• Stopword list: Common English function words, generic adjectives, and overly broad nouns are excluded from single-word matches
• Multi-word phrases bypass the stopword filter (they're more likely to be actual concepts)

CLI tool

• Globally installable via npm install -g wiki-referencify
• Binary: wiki-referencify [options] [file]
• Reads from stdin if no file is provided
• Uses lino-arguments for option parsing with environment variable support

Usage: wiki-referencify [options] [file]

Options:
  --auto-disambiguation  Auto-resolve disambiguation using context  [boolean]
  --max-phrase-length    Max words to consider as a single phrase  [number]
  -v, --verbose          Enable verbose logging to stderr           [boolean]
  -h, --help             Show help                                  [boolean]

Example

Input:

# Introduction to Machine Learning

Machine learning is a branch of artificial intelligence. Python is
a popular programming language for machine learning.

Output:

# Introduction to Machine Learning

[Machine learning](https://en.wikipedia.org/wiki/Machine_learning) is a
[branch](https://en.wikipedia.org/wiki/branch_(disambiguation)) of
[artificial intelligence](https://en.wikipedia.org/wiki/artificial_intelligence_(disambiguation)).
[Python](https://en.wikipedia.org/wiki/Python) is a popular
[programming language](https://en.wikipedia.org/wiki/Programming_language) for
[machine learning](https://en.wikipedia.org/wiki/Machine_learning).

Architecture

src/
  cli.js              # CLI entry point using lino-arguments
  index.js            # Module exports
  index.d.ts          # TypeScript type definitions
  referencify.js      # Core algorithm orchestration
  markdown-processor.js  # Markdown parsing, skip regions, n-gram generation
  wikipedia-api.js    # Wikipedia API client with caching

Testing

• 26 unit tests covering all core modules
• Tests work on Node.js, Bun, and Deno (via test-anywhere framework)
• All CI checks pass: ESLint, Prettier, jscpd duplication check

Test plan

• Run npm test — all 26 tests pass
• Run npm run check — lint, format, duplication all pass
• Test CLI with stdin: echo "# Test\n\nPython is a programming language." | wiki-referencify
• Test CLI with file argument: wiki-referencify examples/basic-usage.js
• Verify headers are not linked
• Verify existing links are not re-linked
• Verify longer phrases are preferred over single words
• Verify disambiguation URLs are used when disambiguation pages exist

---

🤖 Generated with Claude Code

[konard]

🤖 Solution Draft Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Public pricing estimate: $8.423931
• Calculated by Anthropic: $6.704218 USD
• Difference: $-1.719712 (-20.41%)
  📎 Log file uploaded as Gist (5249KB)
  🔗 View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

🔄 Auto-restart 1/3

Detected uncommitted changes from previous run. Starting new session to review and commit them.

Uncommitted files:

?? issue-details.txt
?? lint-fix-results.txt
?? lint-results.txt
?? lint-results2.txt
?? test-results.txt

---

Auto-restart will stop after changes are committed or after 2 more iterations. Please wait until working session will end and give your feedback.

[konard]

🔄 Auto-restart 1/3 Log

This log file contains the complete execution trace of the AI solution draft process.

💰 Cost estimation:

• Public pricing estimate: $0.324616
• Calculated by Anthropic: $0.246044 USD
• Difference: $-0.078572 (-24.20%)
  📎 Log file uploaded as Gist (5825KB)
  🔗 View complete solution draft log

---

Now working session is ended, feel free to review and add any feedback on the solution draft.

[konard]

✅ Ready to merge

This pull request is now ready to be merged:

• All CI checks have passed
• No merge conflicts
• No pending changes

---

Monitored by hive-mind with --auto-restart-until-mergeable flag