Introducing AetherShell, Simon, and Chasm

Today, we are excited to announce three new open-source tools from NERVOSYS, each built in Rust and designed from the ground up for the age of AI agents. Together they form a powerful triad: AetherShell, Simon, and Chasm.

AetherShell — The AI-Native Command Line

Traditional shells were designed for humans typing text commands. AetherShell reimagines the command line for a world where AI agents are first-class participants.

AI agents today are forced to interact with shells designed decades ago for humans — Bash on Linux, PowerShell on Windows, Zsh on macOS. Each has its own syntax, conventions, and quirks. Worse, their output is unstructured text that varies across OS versions, locales, and tool installations, creating non-deterministic results that make agent workflows fundamentally brittle.

An agent that parses df -h output on Ubuntu will break on Alpine. A script that scrapes Get-Process output will fail when column widths change. Every shell command becomes a fragile integration point, and agents spend more time wrestling with output formats than solving real problems.

AetherShell eliminates this entire class of failure. It provides a single cross-platform shell language with deterministic, typed output — every command returns structured data, not raw text. An ontology built into the shell makes commands, their arguments, and their return types discoverable by AI agents without documentation scraping or prompt engineering. One language, every platform, predictable results.

Instead of piping raw text between programs, AetherShell uses typed pipelines — structured data flows with compile-time validation. You can chain operations like where, map, sort_by, and to_json on rich objects, not just strings.

The real breakthrough is the agent system. Spawn autonomous AI agents directly from your shell with a system prompt, model selection (OpenAI, Anthropic, and more), and fine-grained tool access to the filesystem, HTTP, and system information. Agents can investigate, reason, and act independently.

Need more power? AetherShell supports swarm orchestration — deploy multi-agent swarms with configurable strategies like divide-and-conquer for parallel analysis tasks such as codebase security audits.

Additional capabilities include a DAG-based workflow engine with automatic parallelization, Git-native version control for pipelines, and sandboxed execution environments for secure agent operation.

Install it today: cargo install aethershell

Simon — Agentic System Monitoring

Simon (Silicon Monitor) is the world's first agentic system monitoring utility. While traditional monitoring tools produce dashboards for humans to read, Simon exposes deep hardware telemetry to AI agents via natural language queries and the Model Context Protocol (MCP).

Ask your AI assistant "What is using the most GPU memory?" or "Which disk is closest to failure?" and get precise, actionable answers grounded in real telemetry data.

Under the hood, Simon provides comprehensive metrics across every hardware domain:

• GPU — NVIDIA (NVML), AMD (ROCm/sysfs), and Intel Arc. Utilization, VRAM, temperature, power draw, fan speed, clocks, and per-process GPU memory attribution.
• CPU & Thermal — Per-core metrics, frequencies, temperatures, and hybrid architecture awareness for Intel P/E cores and ARM big.LITTLE.
• Disk & Storage — I/O throughput, queue depth, and SMART health data for predictive failure detection.
• Network — Interface statistics, bandwidth monitoring, plus built-in nmap-style port scanning, ping, traceroute, and DNS lookup.
• Peripherals — Audio devices, Bluetooth, displays, and full USB topology.

Simon ships with four interfaces: AI agent mode (simon ai), CLI (simon cpu), TUI (simon tui), and GUI (simon gui). All output is available as structured JSON for scripting and automation.

Install it today: cargo install simon

Chasm — Bridging the Divide Between AI Providers

If you use more than one AI assistant — and most developers now do — your conversations are scattered across VS Code Copilot, Cursor, Claude, ChatGPT, Gemini, Ollama, and dozens of other tools. Chasm solves this.

Chasm is a local-first tool that harvests, merges, and search-indexes AI chat sessions from 30+ providers and editors into a single unified SQLite database with full-text search. Your data never leaves your machine.

Key capabilities include:

• Workspace Discovery — Automatically scans for VS Code, Cursor, Windsurf, and Claude Code workspaces. Recovers orphaned sessions from old workspace hashes and crashes.
• Agent Launcher — Launch terminal AI agents (Aider, Claude Code, Codex, Gemini, Goose, Amp, Cline) with automatic session capture on exit.
• Watch Mode — Real-time filesystem monitoring with debounced auto-harvest.
• Share Link Import — Import shared conversations from ChatGPT, Claude, and Gemini via URL.
• MCP Server — Let your AI agents query your conversation history directly.

Chasm deploys everywhere: CLI, Tauri desktop app, React Native mobile, web dashboard, VS Code extension, browser extension, and plugins for JetBrains and Vim/Neovim.

For enterprise teams, Chasm offers AES-256-GCM encryption at rest, SSO/SAML, RBAC team workspaces, audit logging, and SOC2/HIPAA/GDPR compliance.

Install it today: cargo install chasm

A Unified Vision

AetherShell, Simon, and Chasm share a common architecture: designed for use by agents, built in Rust for performance and safety, CLI-first with rich alternative interfaces, and deeply integrated with the Model Context Protocol (MCP) and ontologies for seamless AI agent interoperability and discoverability.

They represent our belief that the future of developer tooling is agentic — tools that work equally well for humans and AI agents, with structured data, natural language interfaces, and local-first principles.

All three are open source and available on GitHub. We'd love your feedback!

Per aspera ad astra 🚀