AGI CLI

AGI CLI (agi-cli / agents-cli) is an open-source terminal agent harness and orchestration framework developed by phnx-labs. It provides a modular command-line infrastructure designed to run, manage, and coordinate AI coding agents directly within developer workspaces. In day-to-day operation, AGI CLI coordinates multi-agent routines and tool invocation through a structured CLI environment. The harness leverages project-level configuration and context files (such as CLAUDE.md and GEMINI.md) to integrate with major foundation models including Anthropic Claude and Google Gemini, delegating specific coding and automation tasks across repository workflows. AGI CLI differentiates itself through a modular skills architecture and dedicated configuration surfaces defined under open specifications. Its codebase incorporates native bindings, distinct agent definitions, and reusable skills modules that decouple task execution logic from the underlying LLM harness. The project is fully open source under the phnx-labs organization on GitHub, distributed as a monorepo containing modular npm packages and native CLI tooling.

Tags: agent harness, ai agent, cli, developer tools, OpenSource

Visit AGI CLI

Key features

  • Modular Skills System — Extensible tooling engine allowing custom skills and CLI commands to be organized into dedicated modules. - Multi-Provider LLM Integration — Configured for interaction with leading models including Anthropic Claude and Google Gemini via dedicated specification files. - Multi-Agent Orchestration — Built-in harness to manage discrete agent definitions and workflows within the .agents directory structure. - Open Configuration Specs — Uses OpenSpec configuration surface standards for predictable agent behaviors and flags. - Native Components — Incorporates native runtime packages for low-overhead terminal execution and tool invocation. - Open-Source Monorepo — Maintained as an open TypeScript/native monorepo with modular package distribution via npm.

Use cases

Use Case: Automated CLI Agent Workflows Developers run multi-step code generation, testing, and file editing routines triggered and supervised through the command line without switching to a web or IDE interface.

Use Case: Custom Terminal Skills Development Engineers define project-specific tooling and automated scripts inside the skills/ directory to give terminal agents direct access to domain-specific CLI utilities and internal APIs.

Use Case: Multi-Model Agent Harnessing Teams configure and benchmark different models (such as Claude and Gemini) against repository tasks using structured contextual guidelines like CLAUDE.md and GEMINI.md.

Who it is for

AGI CLI is built for developers, tool builders, and platform engineers who want a customizable, code-first harness for orchestrating terminal AI agents with modular skills. It is less suitable for non-technical users or developers looking for a turnkey, closed-source desktop GUI agent.

Install guide

Installation Install the package globally via npm (standard distribution based on repository structure) or clone the repository to build locally: bash # Global installation via npm npm install -g @phnx-labs/agi-cli # Or clone and build from source git clone https://github.com/phnx-labs/agi-cli.git cd agi-cli npm install npm run build

First Run Export your desired model API keys in your shell environment: bash export ANTHROPIC_API_KEY="your-anthropic-key" export GEMINI_API_KEY="your-gemini-api-key" Initialize or verify the CLI configuration within your project: bash agi --version agi init

Example Session Run an agent task to inspect the repository and run a modular skill: bash agi run "Audit the codebase for missing tests and generate unit tests for untested packages"

Part of CLI Coding Agents

Categories: AI Coding Agents, Code Assistants, Developer Tools

Related CLI agents

  • SWE-agent — SWE-agent is an open-source autonomous software engineering agent harness built and maintained by researchers from Princeton University and Stanford University.…
  • Agentic Engineering Framework — Agentic Engineering Framework is an open-source orchestration harness and governance layer developed by Dimitri Geelen. It is designed to constrain,…
  • Hivelore — Hivelore is an open-source agent harness and contextual memory orchestrator created by Doucs91. It provides persistent project memory, decision tracking,…
  • Weaver — Weaver is an open-source agent harness and orchestration framework maintained by sean35mm on GitHub. Built on top of the Bun…
  • Stratless — Stratless is an open-source agent harness and execution orchestrator developed by the Stratless AI team. It provides a modular infrastructure…
  • Context Bridge — Context Bridge is an open-source CLI agent harness and context orchestration utility developed by serdardb. It is built to facilitate…
  • AI Tools
  • Categories
  • Industries
  • CLI Coding Agents
  • MCP Servers
  • MCP Categories