rtk-optimized.md

title	description	tags
RTK Token Optimization Template	CLAUDE.md configuration for minimizing token consumption with RTK CLI proxy	claude-md template performance

RTK Token Optimization

Context: Using RTK (Rust Token Killer) to minimize token consumption from command outputs.

Commands to Optimize

Always use RTK wrapper for these high-verbosity commands:

Git Operations (92.3% avg reduction)

• rtk git log instead of git log
• rtk git status instead of git status
• rtk git diff instead of git diff

File Operations (69.4% avg reduction)

• rtk find "*.md" . instead of find . -name "*.md"
• rtk read <file> instead of cat <file> (for large files >10K lines)
• rtk ls . instead of ls -la
• rtk grep "pattern" instead of grep -r "pattern"

JS/TS Stack (70-90% reduction)

• rtk vitest run instead of pnpm test
• rtk pnpm list instead of pnpm list
• rtk pnpm outdated instead of pnpm outdated
• rtk prisma migrate status instead of pnpm prisma migrate status

Rust Toolchain (80-90% reduction)

• rtk cargo test instead of cargo test
• rtk cargo build instead of cargo build
• rtk cargo clippy instead of cargo clippy

Python (90% reduction)

• rtk python pytest instead of pytest

Go (90% reduction)

• rtk go test instead of go test

GitHub CLI (79-87% reduction)

• rtk gh pr view <num> instead of gh pr view <num>
• rtk gh pr checks <num> instead of gh pr checks <num>

Token Savings Target

Baseline: ~150K tokens per 30-min session With RTK: ~45K tokens (70% reduction)

Installation

# Homebrew (macOS/Linux)
brew install rtk-ai/tap/rtk

# Cargo (all platforms)
cargo install rtk

# Hook-first install
rtk init

Verification

Check RTK availability:

rtk --version  # Should show: rtk 0.16.0+

When NOT to use RTK

• Quick exploration (1-2 commands): overhead not worth it
• Already using tools like Grep/Read (Claude native tools are optimized)
• Small outputs (<100 chars): minimal gain

Automation

Use RTK automatically via hook (see .claude/hooks/bash/rtk-wrapper.sh) or rtk init for hook-first setup.