xerxes 0.3.1

An intelligent agent with BYOK for managing cloud infrastructure & bash env

Xerxes

An intelligent DevOps agent with unrestricted CLI access and free-form reasoning

Features • Installation • Quick Start • Documentation • Contributing

---

Overview

Xerxes is an intelligent DevOps assistant that transforms natural language into bash commands. Unlike traditional tools with predefined commands, Xerxes leverages full bash capabilities with intelligent resource discovery to execute complex operations across your entire infrastructure.

Features

Core Capabilities

• Full Bash Shell Access: Execute ANY command with complete shell features
  • Pipes, redirection, chaining (|, >, >>, &&, ||)
  • Variables, loops, conditionals, subshells
  • Process substitution and command expansion
• Intelligent Resource Discovery: Automatic fuzzy matching and pattern-based discovery
  • Never assumes exact names - always discovers first
  • Pattern: SEARCH → MATCH → OPERATE
  • Works with files, pods, containers, instances, and more
• Token-Efficient Operations: Minimal output formatting for faster responses
  • Uses -o name, --format, --query flags
  • Field projection to reduce token usage
• Interactive Approval: See command + reasoning before execution
  • [R]un - Execute this command
  • [S]kip - Skip and continue
  • [A]lways - Auto-approve for session
• Multi-Cloud & Multi-Tool: kubectl, docker, aws, gcloud, helm, jq, ffmpeg, git, and more
• Safety First: Automatic detection of destructive operations with confirmation prompts

Installation

Using UV (Recommended)

uv tool install xerxes

To update to the latest version:

uv tool upgrade xerxes

From Source

git clone https://github.com/shammianand/xerxes.git
cd xerxes
uv sync

Quick Start

1. Prerequisites

• Python 3.10+
• GCP Project with Vertex AI API enabled
• Service Account or API key for Vertex AI
• Optional: Install CLI tools you want to use (kubectl, docker, aws, gcloud)

2. Configuration

Create a .env file:

XERXES_VERTEX_PROJECT_ID=your-gcp-project-id
XERXES_VERTEX_LOCATION=us-central1
XERXES_VERTEX_MODEL=gemini-2.0-flash
XERXES_GOOGLE_APPLICATION_CREDENTIALS=/path/to/service-account.json

Or use the CLI:

xerxes config set vertex_project_id your-gcp-project-id
xerxes config set google_application_credentials /path/to/sa.json

3. Start Chatting

xerxes chat

Usage

Example Session

https://github.com/user-attachments/assets/350d91eb-b999-4c18-aba3-52d4e430bb8e

Operations:

"Scale the api deployment to 5 replicas"
"Delete all pods in failed state"
"Create a new GCS bucket for backups"

CLI Commands

# Start interactive chat
xerxes chat

# Manage configuration
xerxes config show
xerxes config set <key> <value>

# List available tools
xerxes tools

# Show version
xerxes version

Configuration Options

Variable	Description	Default
XERXES_VERTEX_PROJECT_ID	GCP project ID (required)	-
XERXES_VERTEX_LOCATION	GCP region	us-central1
XERXES_VERTEX_MODEL	Model name	claude-3-5-sonnet@20240620
XERXES_GOOGLE_APPLICATION_CREDENTIALS	Path to service account JSON	-
XERXES_MAX_TOKENS	Max tokens per response	4096
XERXES_TEMPERATURE	LLM temperature	0.0

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes
4. Run tests and linters:

   uv run pytest
   uv run black src/
   uv run ruff check src/
   

5. Commit your changes (git commit -m 'Add amazing feature')
6. Push to the branch (git push origin feature/amazing-feature)
7. Open a Pull Request

Development Setup

# Clone the repo
git clone https://github.com/shammianand/xerxes.git
cd xerxes

# Install dependencies
uv sync --all-extras

# Run in development mode
uv run xerxes chat

Roadmap

• Additional LLM providers (Anthropic direct, OpenAI, Ollama)
• Terraform integration
• Ansible playbook execution
• Multi-step workflow automation
• Session history save/replay
• Web dashboard for monitoring
• Plugin system for custom tools
• One-shot command mode (xerxes run "list all pods")

License

MIT License - see LICENSE for details

Support

• Bug Reports: Open an issue
• Feature Requests: Start a discussion
• Documentation: Coming soon

---

Built with ❤️ by Shammi Anand