Summary

Optimizes token consumption for structured and markdown data:

1. Null filtering: filtered out null values for web_data_* tools, reducing tokens by 5-10%

2. Markdown Minification: Implements remark + strip-markdown plugin for scrape_as_markdown tool, achieving 60% token reduction by removing base64 images, HTML tags, and formatting while preserving meaningful content

3. Google Search Response Sanitization: Cleanses search results by normalizing whitespace, removing Unicode formatting characters, deduplicating related keywords, and extracting only essential fields (link, title, description) from organic results (Thanks to Nikita for this valuable feedback | example response)

Changes

• Added remark/strip-markdown processing pipeline for markdown responses
• Implemented clean_google_search_payload() to sanitize and normalize search engine results

(Still being tested and evaluated for efficiency)