OneCLI is an open-source AI agent harness and execution gateway created by OneCLI (a Y Combinator-backed company) designed to safely deploy and orchestrate autonomous agents across engineering teams. The platform acts as a secure intermediary between AI models and underlying infrastructure, enabling agents to execute tasks while enforcing deterministic operational boundaries. In daily operations, team members interact with agents via chat interfaces such as Slack or through command interfaces. Agents operate within sealed sandboxes and perform automated tasks—such as diagnosing error spikes, opening pull requests, and querying databases—while all API requests and tool invocations are routed through the OneCLI gateway. The gateway dynamically injects scoped credentials per request, ensuring the LLM itself never has direct access to raw secrets or passwords. OneCLI differentiates itself from traditional CLI agents by enforcing security and governance outside the model's context window. Instead of relying on prompt instructions to prevent unintended behavior, OneCLI enforces hard rules at the proxy and network level, including rate limiting against runaway loops, explicit blocks on destructive commands, and mandatory human-in-the-loop approval prompts for sensitive operations. OneCLI is open-source software with source code available on GitHub. Hosted and team management services are provided via app.onecli.sh, offering a free tier that includes 500 calls per month and $5 in AI credits.
Zero-Secret Sandbox Execution — Injects scoped credentials per request through a secure gateway so models never hold raw API keys or passwords. - External Deterministic Guardrails — Enforces execution constraints at the proxy level rather than relying on system prompt adherence. - Human-in-the-Loop Approvals — Automatically pauses agent execution and requests interactive authorization before executing sensitive actions. - Runaway Loop Protection — Monitors call frequency and automatically throttles or halts agents that exhibit repetitive or runaway behavior. - Role-Based Tool Scoping — Maps agent capabilities strictly to the permissions and tooling access of the specific user running the agent. - Team Collaboration Integration — Surfaces agent updates, task completions, and interactive approval cards directly within workspace channels like Slack.
Use cases
Use Case: Automated Error Triage and PR Generation An agent monitors overnight application logs, traces a sudden spike in errors back to a recent commit, and opens a GitHub pull request containing a suggested fix, notifying the on-call engineer for review.
Use Case: Safe Database Auditing with Action Gates A developer instructs an agent to cross-reference payment records from Stripe against a PostgreSQL database. The agent processes thousands of transactions, aggregates mismatches, and pauses automatically for human sign-off before initiating high-value refunds.
Use Case: Secure Operational Offboarding An internal operations agent automates contractor account deprovisioning across multiple developer services via the OneCLI gateway without exposing administrative credentials to the underlying LLM context.
Who it is for
OneCLI is designed for software engineering teams, platform administrators, and technical organizations that want to deploy autonomous AI agents without exposing production credentials or risking unconstrained system actions. It is less suited for solo developers seeking a lightweight, standalone terminal coding assistant that does not require team governance, proxy infrastructure, or approval workflows.
Install guide
Installation OneCLI can be accessed via the web application or set up locally via its open-source repository. To install the CLI client via npm (standard installation method): bash npm install -g @onecli/cli Alternatively, clone and inspect the open-source harness from GitHub: bash git clone https://github.com/onecli/onecli.git cd onecli
First Run 1. Create an account at https://app.onecli.sh/ to configure your organization and gateway rules. 2. Authenticate the CLI with your OneCLI workspace: bash onecli auth login 3. Connect your desired tool integrations (such as GitHub, Postgres, or Stripe) via the OneCLI dashboard to configure scoped credential injection.
Example Session Execute a sandboxed agent task requiring database verification: bash onecli run "Audit yesterday's failed webhook deliveries in PostgreSQL and draft a summary of impacted accounts" The agent runs inside its isolated sandbox, accesses PostgreSQL via the gateway proxy without reading raw database passwords, and prompts for confirmation if any mutating query or external notification is triggered.
nanoclaw — nanoclaw is an open-source CLI agent framework and containerized execution harness developed by Nanoco AI. Designed around Claude-powered workflows and…
AionUi — AionUi is an open-source desktop multi-agent harness and orchestrator developed by iOfficeAI. It serves as a unified workspace for managing,…
Orkas — Orkas is an open-source desktop agent harness and workflow orchestrator developed by Orkas-AI. Built to bridge autonomous AI execution with…
Nullclaw — Nullclaw is an open-source, lightweight AI coding assistant and agent harness designed to operate with minimal overhead. Developed using Zig…
Hermes Agent — Hermes Agent is an open-source terminal-based AI coding agent developed by NousResearch. Designed to bridge terminal workflows and modern IDE…
SWE-agent — SWE-agent is an open-source autonomous software engineering agent harness built and maintained by researchers from Princeton University and Stanford University.…