DevOps CLI Toolset

A robust Rust-based CLI toolset designed for efficient DevOps workflows.

繁體中文 | 简体中文 | 日本語

Quick Overview

Feature	Description
Terraform Cleanup	Remove .terraform, .terragrunt-cache and lock files
AI Assistant Upgrade	Batch update Claude Code, Codex, Gemini CLI
Package Manager	Install/update nvm, pnpm, Rust, Go, kubectl, k9s, tmux, vim, etc.
MCP Manager	Manage MCP servers for Claude/Codex/Gemini
Security Scanner	Run gitleaks, trufflehog, git-secrets, trivy, semgrep
Prompt Generator	4-step LLM workflow with progress tracking
Skill Installer	Install AI CLI extensions (Claude/Codex/Gemini)
Rust Build	Build cross-platform binaries (cargo/cross)
Container Builder	Docker/Buildah multi-arch builds (x86, arm64, armv7, Jetson)
Kubeconfig Manager	tmux window-isolated kubeconfig

Features

1. Terraform/Terragrunt Cache Cleaner

Intelligently cleans cache files generated by Terraform and Terragrunt:

• .terragrunt-cache directories
• .terraform directories
• .terraform.lock.hcl files
• Automatically filters duplicate sub-paths to avoid redundant deletions.

2. AI Code Assistant Upgrader

Batch upgrades for popular AI code assistants:

• Claude Code (@anthropic-ai/claude-code)
• OpenAI Codex (@openai/codex)
• Google Gemini CLI (@google/gemini-cli)

3. Package Installer/Updater (macOS/Linux)

Install, remove, and update common tools with an interactive checklist:

• nvm (installs latest Node.js)
• pnpm
• Rust (via rustup)
• Go (downloads latest official archive/pkg)
• Terraform
• kubectl
• kubectx
• k9s
• git
• uv (installs latest Python)
• tmux (includes TPM + tmux.conf setup)
• vim (includes vim-plug + molokai config)
• ffmpeg (build script on Linux, Homebrew on macOS)

4. MCP Tool Management

Manages MCP servers for Claude, Codex, and Gemini CLI:

MCP Tool	Description
sequential-thinking	Sequential Thinking
context7	Documentation Query
chrome-devtools	Chrome DevTools
kubernetes	K8s Management
tailwindcss	Tailwind CSS (Utility-first CSS)
arxiv-mcp-server	arXiv Academic Paper Search & Download
github	GitHub Integration
cloudflare-*	Cloudflare MCP Servers (Docs/Workers/Observability/etc.)

Optional Feature Flags & MCP Credentials (build-time only via .env):

How to enable:

1. cp .env.example .env
2. Fill in the values you need.
3. Build the binary (cargo build --release).

Available MCP options:

• Context7 MCP: set CONTEXT7_API_KEY to enable the context7 tool.
• GitHub MCP: set GITHUB_PERSONAL_ACCESS_TOKEN (required) to enable the github tool.
  • GITHUB_MCP_MODE: docker (default, works without Copilot subscription) or remote (requires GitHub Copilot subscription)
  • GITHUB_HOST: optional, default github.com (for GitHub Enterprise)
  • GITHUB_TOOLSETS: optional, comma-separated list of toolsets to enable (e.g., repos,issues,pull_requests,actions)
• Cloudflare MCP: set enable_cloudflare_mcp=true to enable cloudflare-* tools (OAuth during install).
• arXiv MCP: set ARXIV_STORAGE_PATH (optional, default ~/.arxiv-papers) to configure paper storage location.

For Codex MCP installs, build-time values for CONTEXT7_API_KEY, GITHUB_PERSONAL_ACCESS_TOKEN, and GITHUB_HOST are written into ~/.codex/config.toml, so runtime env vars aren't required. Cloudflare MCP installs use OAuth in interactive mode (stdio inherited) so URL prompts and auth codes work. On WSL, open the auth URL with wslview or run the CLI on Windows if the localhost callback cannot reach WSL. Cloudflare's MCP server catalog: https://developers.cloudflare.com/agents/model-context-protocol/mcp-servers-for-cloudflare/

5. Project Security Scanner

Quickly installs and runs strict scans against the current Git repo:

• gitleaks (history + working tree)
• trufflehog (history + working tree)
• git-secrets (history + working tree)
• trivy (SCA + misconfig on working tree)
• semgrep (SAST on working tree)

Auto-install tries common package managers, Trivy's install script, and pipx/venv for Semgrep, then falls back to GitHub releases (requires curl/wget and tar/unzip). Working tree scans only include Git-tracked files and skip anything ignored by .gitignore, with raw logs printed per scan.

6. LLM Prompt Generator

Generate and execute 4-step prompts for LLM workflows:

• Generate: Create prompt files from YAML/JSON specification files
• Run: Interactive runner for Claude/Codex/Gemini CLI (run all or selected features)
• Status: View feature execution progress and status
• Validate: Validate spec file format
• YAML Prompt: Generate a YAML spec prompt (built-in template)
• Spec field verification_url may be empty to indicate no validation URL

The 4-step workflow:

1. P1: Requirements, implementation, and deployment
2. P2: E2E validation in verification environment
3. P3: Refactoring and optimization
4. P4: E2E regression testing in verification environment

Progress is tracked per feature with session management for resumable execution.

7. Skill Installer

Install and manage extensions for AI CLI tools:

CLI	Extension Format	Install Path
Claude Code	Plugins + Skills	~/.claude/plugins/, ~/.claude/skills/
OpenAI Codex	Skills (SKILL.md)	~/.codex/skills/
Google Gemini	Extensions (TOML)	~/.gemini/extensions/

Available Extensions:

• ralph-wiggum - AI Agent Loop (Claude/Gemini)
• security-guidance - Security Best Practices (Claude/Gemini)
• frontend-design - Frontend Interface Design (All CLIs)
• code-review - Code Review Assistant (All CLIs)
• pr-review-toolkit - PR Review Tools (All CLIs)
• commit-commands - Git Commit Helper (All CLIs)
• writing-rules - Writing Style Rules (All CLIs)

Note: Gemini uses a different extension format. The installer automatically converts Claude plugins to Gemini's native TOML format and registers them in extension-enablement.json.

See docs/SKILL_INSTALLER.md for development guide.

8. Container Image Builder

Build multi-architecture container images with Docker or Buildah:

• Build Engines: Docker (buildx) or Buildah (daemonless OCI builder)
• Multi-Architecture Support:
  • x86_64 / amd64 (Intel/AMD 64-bit)
  • arm64 / aarch64 (Apple Silicon, AWS Graviton)
  • armv7 / arm/v7 (Raspberry Pi 2/3)
  • Jetson Nano (NVIDIA Jetson Nano aarch64)
• Dockerfile Scanner: Auto-detects Dockerfile, Containerfile, and variants (Dockerfile.dev, etc.)
• Registry Push: Optional push to container registries
• Quick Selection: Remembers recent image names, tags, and registries for fast reuse

Installation

via Install Script (Recommended - Linux/macOS)

curl -fsSL https://raw.githubusercontent.com/DennySORA/Ops-Tools/main/install.sh | bash

Manual Installation

Download the version corresponding to your system from the Releases page.

Build from Source

# Build
cargo build --release

# Configure environment variables (Optional, for MCP management features)
cp .env.example .env
# Edit .env and fill in your credentials

Usage

If installed via the script, run directly:

ops-tools

If built from source:

cargo run
# or
./target/release/tools

Select a function from the menu:

1. Clean Terraform/Terragrunt cache files
2. Upgrade AI code assistant tools
3. Install/Update system packages (macOS/Linux)
4. Upgrade Rust projects and toolchain
5. Build Rust binaries for multiple platforms (glibc vs musl, cargo or cross)
6. Security scan (Gitleaks/TruffleHog/Git-Secrets/Trivy/Semgrep)
7. Manage MCP tools (Claude/Codex/Gemini)
8. LLM Prompt Generator (4-step workflow)
9. Container Image Builder (Docker/Buildah multi-arch)
10. Kubeconfig Manager (tmux window-isolated)
11. Language settings (English/Traditional Chinese/Simplified Chinese/Japanese)

Rust Build target hints

• *-unknown-linux-gnu (glibc): best for mainstream distros; dynamic linking, smaller binary but needs system glibc.
• *-unknown-linux-musl (musl, mostly static): best for Alpine/scratch images; single-file style deployment.
• i686-* legacy 32-bit x86; powerpc64le-* OpenPOWER/IBM Cloud; wasm32-unknown-unknown for browser/wasm runtimes (no std).

You'll be prompted to choose a language at startup, and you can switch languages later from the menu. Language preference is saved to ~/.config/ops-tools/config.toml (Linux), ~/Library/Application Support/ops-tools/config.toml (macOS), or %APPDATA%\\ops-tools\\config.toml (Windows).

Contributing

Contributions are welcome! Please submit a Pull Request or open an Issue.

License

MIT License