AI Gentle Stack configures and enhances existing AI coding agents by integrating a comprehensive "Gentleman stack." It addresses the limitations of standalone agents by providing persistent memory, structured Spec-Driven Development (SDD) workflows, curated coding skills, and real-time documentation access. This ecosystem empowers developers, researchers, and power users to leverage their chosen AI agents more effectively, transforming them from simple chatbots into sophisticated coding assistants that can also teach.

How It Works

The project acts as an ecosystem configurator, not an agent installer. It injects the "Gentleman stack" into popular AI coding agents like Claude Code, OpenCode, Gemini CLI, Cursor, and VS Code Copilot. Core components include Engram for cross-session memory, an SDD workflow for structured planning and execution, a library of curated coding Skills, Context7 for real-time documentation lookup (MCP servers), a security-first Permissions system, and a teaching-oriented Persona. This approach unifies advanced capabilities across diverse AI tools.

Quick Start & Requirements

Installation is streamlined via platform-specific scripts: macOS/Linux use a curl bash script, Windows uses a PowerShell irm script. Alternatively, Homebrew (brew install gentle-ai) or Go 1.24+ (go install github.com/gentleman-programming/gentle-ai/cmd/gentle-ai@latest) can be used. The primary interface is an interactive TUI. No Go toolchain is required for the script-based installations. Docker is needed for running end-to-end tests.

Highlighted Details

• Supported Agents: Claude Code, OpenCode, Gemini CLI, Cursor, VS Code Copilot.
• Core Components: Engram (persistent memory), SDD (9-phase Spec-Driven Development workflow), Skills (curated coding patterns), Context7 (MCP server for live docs), Persona (teaching-oriented), Permissions (security-first), GGA (AI provider switcher).
• Skills Library: Includes SDD phases (e.g., sdd-spec, sdd-apply) and foundational skills like go-testing and skill-creator.
• Presets: Offers full-gentleman, ecosystem-only, minimal, and custom configurations.
• Persona Modes: gentleman (mentoring), neutral, and custom.
• Cross-Platform: Binaries available for Linux, macOS, and Windows (amd64/arm64), supporting apt, pacman, brew, and winget package managers.

Maintenance & Community

The provided README does not contain specific details regarding notable contributors, sponsorships, or community channels (e.g., Discord, Slack).

Licensing & Compatibility

The project is released under the MIT license, which is highly permissive and allows for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

Sub-agent functionality for Gemini CLI is marked as experimental. Windows users may encounter heuristic antivirus flags on unsigned Go binaries, necessitating manual checksum verification against release artifacts. The README does not detail specific performance benchmarks or known bugs.