Optimize memory usage, improve YAML parsing, and add CLI enhancements

[Binb1]

Summary

This PR improves performance and reliability across multiple areas: streaming large JSONL files to reduce memory footprint, enhancing the YAML parser to support multi-line strings, refining CLI argument parsing with short flags, and adding comprehensive test coverage.

Key Changes

Memory & Performance

• Replace readFileSync with streaming in stats.ts to handle large session files without loading entire contents into memory
• Implement line-by-line processing with proper chunk boundary handling
• Add caching to findProjectRoot() to avoid repeated filesystem traversal
• Improve token estimation algorithm from simple length-based to word+symbol heuristic for better accuracy

YAML Parser Enhancement

• Add support for multi-line literal (|) and folded (>) blocks in skill frontmatter
• Implement proper indentation tracking and block termination detection
• Refactor parser with helper functions (flushMultiLine, flushList) for cleaner state management
• Add comprehensive documentation of parser capabilities and limitations

CLI Improvements

• Add short flag aliases: -j (json), -a (agent), -t (type), -s (scope), -p (period), -v (version), -h (help)
• Support combined short flags (e.g., -js) with proper boolean/value handling
• Support --key=value syntax for long flags
• Distinguish between boolean-only flags and value-accepting flags

Dry-Run Support

• Add --dry-run flag to grab and rm commands for preview mode
• Update output formatting to indicate dry-run status with visual indicators
• Include dryRun flag in JSON output responses

GitHub Fetch Robustness

• Replace simple response.text() with streaming reader to enforce size limits regardless of content-length header
• Properly handle missing or incorrect content-length headers
• Implement byte-by-byte size checking with early cancellation

Path Cleaning Logic

• Refactor cleanDirName() with two-phase approach: strip HOME prefix segments in order, then skip container directories
• Improve accuracy of project name extraction from encoded directory paths

Testing

• Add comprehensive test suites for parser, health badges, tokens, and GitHub utilities
• Add CI workflow for automated testing on push and pull requests

https://claude.ai/code/session_01879LywhiKaXPsPbcsGgpZi