Neura Market
nono logo

nono

Free

nono - a capability-based, multiplexing sandbox tool, built for developers - lift'n'shift seamless path to prod. Run agents securely without needing any additional infra, zero setup, zero latency.

CybersecurityFreeFree tier
Type
Open Source
Visit Website GitHub (2.2k stars)

About nono

nono is an open-source sandbox for AI agents that provides kernel-level isolation with zero setup and zero latency. Designed for developers, it allows running terminal agents securely without additional infrastructure. Key capabilities include composable JSON policies for fine-grained permissions, tool-specific sandboxing enforcing least privilege, credential protection where secrets are injected at the boundary and zeroised on exit, and a fully auditable cryptographic audit trail. It integrates seamlessly with popular agents like Claude Code, OpenCode, Codex, Antigravity, Goose, OpenClaw, GitHub Copilot, Qwen Code, Pi, and Hermes. nono scales from a single laptop to fleets of agents, and is used in production by companies such as Datadog and Okta.

Key Features

Kernel-level isolation for each agent execution
Zero-latency sandbox with zero setup (curl, Homebrew, or package manager)
Composable JSON policies that version-control alongside code
Tool-specific sandboxing with least privilege by default
Credential protection: secrets injected at boundary, zeroised on exit
Cryptographic, tamper-evident audit trail for every action
Works with multiple terminal agents (Claude Code, OpenCode, Codex, etc.)
Scalable from a single laptop to fleet deployments

Pros & Cons

Pros
  • Fast installation and zero configuration required
  • Fine-grained, composable policies that are easy to manage
  • Per-tool sandboxing reduces blast radius compared to monolithic sandboxes
  • Secrets are never exposed to the agent, enhancing security
  • Fully auditable with a cryptographic trail that can be verified by third parties
  • Actively used by large organizations (Datadog, Okta) in production
Cons
  • Currently limited to terminal-based AI agents, not GUI applications
  • Requires understanding JSON policy syntax for custom configurations

Best For

Running AI agents securely in development and productionSandboxing any terminal-based agent to prevent system compromiseEnforcing per-tool and per-command security policies for agentsManaging and protecting credentials during agent executionAuditing agent behavior with verifiable cryptographic logs

Alternatives to nono

giskard-oss

🐢 Open-Source Evaluation & Testing library for LLM Agents

ai-captcha-bypass

AI Captcha Bypass

backdoor-learning-resources

A list of backdoor learning resources

claude-bug-bounty

AI-powered bug bounty hunting from your terminal - recon, 20 vuln classes, autonomous hunting, and report generation. All inside Claude Code.

agentic-radar

A security scanner for your LLM agentic workflows

FAQ

What is nono?
nono is an open-source sandbox for AI agents that provides kernel-level isolation, zero setup, and a zero-latency execution environment.
How do I install nono?
nono can be installed via curl (curl -fsSL https://nono.sh/install.sh | sh), Homebrew (brew install nono), or using Debian, Fedora, Arch, or Nix packages.
Which AI agents does nono support?
nono supports any terminal agent, including Claude Code, OpenCode, Codex, Antigravity, Goose, OpenClaw, GitHub Copilot, Qwen Code, Pi, Hermes, and more.
How does nono protect credentials?
Secrets are injected at the sandbox boundary and are never exposed to the agent. They are zeroised on exit to prevent leakage.
Is nono free?
Yes, nono is open-source and free to use. It has a free pricing model.