Your team's coding standards, in every AI session.
ACE loads shared skills, tools, and project context from a Git repository called a school. Give your team's Claude Code, Codex, and OpenCode sessions a common starting point, with the skills each project needs.
Works with Claude CodeCodexOpenCode
Built and used at PRODIGY9, a software consultancy that uses ACE across active client projects every day.
one school, every session
Demo: a teacher curates a school of skills and MCP servers and shares it with
git push. Three teammates then run ace setup, launch their backend,
and write different code, all drawing from the same school. Use the step list to replay
any stage.
why teams use it
- Keep coding conventions and review instructions available in every session.
- Update shared skills and tools once instead of maintaining each developer's setup.
- Let each project choose what it needs from the same school.
Built for engineering leads who decide how their team builds software and want those
decisions available wherever the team uses an AI coding tool.
quick start
Latest version: vX.X.X
# macOS / Linux $ curl -fsSL https://ace-rs.dev/install.sh | bash # Windows (PowerShell) > powershell -c "irm https://ace-rs.dev/install.ps1 | iex" # Homebrew (macOS Apple Silicon) $ brew install ace-rs/tap/ace $ ace setup ace-rs/school # clone a school, register MCP, write config $ ace # launch your AI backend
Homebrew installs are managed by brew upgrade ace; ace upgrade is a no-op for brew installs. Intel macOS, Linux, and Windows: use the install scripts above or the release binaries.
build from source
Requires a Rust toolchain (rustup).
# Install latest from the main branch $ cargo install --git https://github.com/ace-rs/ace # Or clone and install from a working tree $ git clone https://github.com/ace-rs/ace $ cd ace $ cargo install --path .
what a school gives you
- Skills and conventions that describe how the team writes and reviews code
- MCP tools, prompts, and project context that ACE supplies to each backend
- Git history for reviewing changes and sharing improvements across projects
how it works
- 1. Install the ACE command-line tool.
- 2. Run ace setup to connect the school, link the selected skills, register its MCP tools, and write project configuration.
- 3. Run ace to load that configuration and launch the chosen AI backend.