Documentation request: Guidance for working with very large files (70k-100k+ records)

Summary

The current documentation lacks guidance for working with very large files - specifically datasets with 70k-100k+ records like parts catalogs, inventory databases, or pricing sheets.

Current State

The docs mention:

• Avoid pasting large content directly
• Use /compact and /clear for context management

But there's no guidance on:

• Practical file size limits
• Chunking strategies
• Alternative approaches (SQLite, indexing, batch processing)
• Workflow examples for large datasets

Proposed Addition

A new documentation page covering:

1. Understanding the Limits

• Context window constraints
• Token consumption for large files
• When direct file reading becomes impractical

2. Practical Strategies

• Query instead of load: Use grep/search to find specific records
• Convert to SQLite: For repeated queries against large datasets
• Split into chunks: Break files by category, range, or size
• Create summary indexes: Lightweight reference files
• Headless mode batching: Process large datasets in batches
• MCP servers: Tools designed for large file handling

3. Examples

• Finding specific parts in a large catalog
• Bulk price updates
• Generating reports from subsets

Why This Matters

Users working with real-world datasets (parts catalogs, product inventories, pricing databases) often have files with 50k-100k+ rows. Without guidance, they'll hit context limits and not know how to proceed.

Draft Content

I've drafted a full documentation page here: https://gist.github.com/ (can provide if helpful)

Would be happy to submit a PR if there's interest.