onecli/onecli

What it solves

OneCLI provides a secure, team-oriented platform for managing autonomous AI agents. It solves the problem of "agent sprawl" in organizations by replacing individual, unmanaged agent setups with a centralized system that handles identity, credential security, and team-wide policy enforcement.

How it works

Each employee is assigned a sandboxed, personal agent that operates within a controlled environment. The platform uses a gateway architecture: agents perform tasks in isolated sandboxes, and when they need to access external services, the gateway intercepts the request to inject necessary credentials or enforce security policies. This ensures that agents never directly handle sensitive keys. The system includes a web dashboard and Slack integration for managing agent memory, skills, and human-in-the-loop approvals for sensitive actions.

Who it’s for

It is designed for companies and teams that want to deploy autonomous agents for employees while maintaining strict control over security, permissions, and access management.

Highlights

• Identity Integration: Provisions agents directly via company identity providers.
• Credential Security: Injects secrets at the gateway level so agents never see or store raw credentials.
• Human-in-the-Loop: Requires explicit approval for sensitive actions like deleting tickets or sending emails.
• Isolated Sandboxes: Every agent runs in its own environment with a dedicated filesystem and shell.
• Flexible Deployment: Supports self-hosting on Kubernetes, laptops, or homelabs with an outbound-only runner that requires no inbound ports.