A CLI tool to safely nuke development and build caches across multiple languages and frameworks.

Join our Discord community for discussions, support, and updates:

• One command to safely clean dev/build caches
• Smart detection of project types and relevant caches
• Safe operations with timestamped backups
• Cross-platform (macOS, Linux, Windows)
• JSON output for scripting and automation
• CI/CD integration with GitHub Actions, GitLab CI, and CircleCI
• Specialized integrations for HuggingFace, PyTorch, Vercel, and Cloudflare
• Advanced NPX analysis with per-package visibility and stale detection
• Enhanced edge cache purging with improved API integration
• System diagnostics with --doctor command

git clone https://github.com/kagehq/cachekill.git
cd cachekill
cargo build --release

cargo install cachekill

# List all caches
./target/release/cachekill --list

# Show what would be cleaned
./target/release/cachekill --dry-run

# Clean with confirmation
./target/release/cachekill

# Clean specific language
./target/release/cachekill --lang js

# Include Docker and NPX
./target/release/cachekill --docker --npx

# JSON output for scripting
./target/release/cachekill --list --json

# System diagnostics
./target/release/cachekill --doctor

# CI mode for automation
./target/release/cachekill --ci prebuild
./target/release/cachekill --ci postbuild

# Specialized integrations
./target/release/cachekill --hf --list
./target/release/cachekill --torch
./target/release/cachekill --vercel --list
./target/release/cachekill --cloudflare

# Advanced NPX analysis with per-package details
./target/release/cachekill --npx --list

# NPX cache management
./target/release/cachekill --npx --dry-run          # Preview what would be cleaned
./target/release/cachekill --npx --force            # Nuclear option - clear all NPX caches
./target/release/cachekill --npx --stale-days 7 --force  # Surgical - only stale packages

• JavaScript/TypeScript: node_modules/, .next/, .vite/
• Python: __pycache__/, .venv/, .pytest_cache/
• Rust: target/, .cargo/
• Java: .gradle/, build/, ~/.m2/repository
• Machine Learning: ~/.cache/huggingface, ~/.cache/torch
• NPX: ~/.npm/_npx
• Docker: Images, containers, volumes

• HuggingFace: Model caches, datasets, and repositories with detailed analysis
• PyTorch: Checkpoints, hub models, and datasets with version tracking
• NPX: Per-package analysis with name, version, size, and stale detection
• Vercel: Enhanced edge cache purging with improved API integration
• Cloudflare: Enhanced edge cache purging with zone-specific targeting

CacheKill provides detailed per-package analysis for NPX caches:

# Analyze NPX packages with detailed breakdown
./target/release/cachekill --npx --list

# Output shows:
# - Package name and version
# - Size and last-used timestamp
# - Stale detection (configurable threshold)
# - Sorted by size (largest first)
# - Summary statistics

Create a .cachekillrc file in your project root:

default_lang = "auto"
stale_days = 14
safe_delete = true
backup_dir = ".cachekill-backup"
exclude_paths = [".git", ".cachekill-backup"]

- uses: ./.github/actions/cachekill
  with:
    mode: 'postbuild'
    args: '--docker --npx'
  env:
    VERCEL_TOKEN: ${{ secrets.VERCEL_TOKEN }}
    CF_API_TOKEN: ${{ secrets.CF_API_TOKEN }}

include:
  - local: 'ci/gitlab-cachekill.yml'

variables:
  CACHEKILL_ARGS: "--docker --npx"

orbs:
  cachekill: cachekill/[email protected]

workflows:
  build:
    jobs:
      - cachekill/prebuild
      - build
      - cachekill/postbuild

• Safe Delete: Moves caches to timestamped backup directory
• Stale Detection: Configurable threshold (default: 14 days)
• Project Detection: Automatically detects project type
• Restore: Use --restore-last to restore from backup

1. Always use --dry-run first to see what will be cleaned
2. Use --list to understand your cache usage before cleaning
3. Configure .cachekillrc for project-specific settings
4. Use --restore-last if a build fails after cleanup

For detailed documentation, see DETAILED.md.

This project is licensed under the FSL-1.1-MIT License. See the LICENSE file for details.

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Run make test
6. Submit a pull request