feat: fully implement 5 research tools with production-ready functionality
Research tools now have complete implementations:
- **page-brief**: Uses @mozilla/readability + jsdom for content extraction,
sbd for sentence parsing, comprehensive error handling with timeout and
network error detection
- **compare-pages**: Uses natural library TF-IDF for text similarity,
detects agreements via high-similarity sentences, identifies conflicts
using negation pattern analysis
- **source-credibility**: Uses tldts for domain parsing, cheerio for HTML
analysis, calculates 6 credibility signals (HTTPS, domain reputation,
author, date, citations, contact info), weighted scoring system
- **claim-checklist**: Uses sbd for sentence boundary detection, regex
patterns for claim identification (statistics, quotes, historical,
scientific, factual), priority levels and evidence suggestions
- **timeline-from-text**: Uses chrono-node for date parsing, sbd for
context extraction, calculates confidence scores, identifies gaps >30 days
All tools follow AI SDK v6 pattern (tool() + jsonSchema()), include proper
TypeScript types, input validation, comprehensive error handling, and
Node.js 18+ fetch requirement verification.
Updated blocks.yml with comprehensive philosophy and domain rules for
quality validation.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
1ea9b83039