Bring Structure to Your Agent Workflows

Building with agents helps you move fast, but practices around reviewing generated code and understanding past sessions are still ad hoc.

The Entire CLI now includes three new commands that bring more structure to that loop: entire recap, entire review, and entire labs.

Look Back at Recent Agent Activity

entire recap is a per-agent breakdown of how you and your team are using AI across your repos. Pick a window with --day (default), --week, --month, or --90 for a rolling 90 days, and switch between solo, team, and side-by-side comparison with --view you|team|both. Filter to a specific agent with --agent.

Once you have signed in with entire login, entire recap opens an interactive TUI by default. Pass --static for plain output suitable for scripts and CI. Each agent gets its own panel covering sessions, checkpoints, token usage, files touched, top skills and MCP servers, and the tool mix behind every run.

Review Your Branch with Multiple Agents

entire review runs the review skills you have configured against the current branch. It is also listed under entire labs while we are still refining the surface, so the flags below may shift. On first run, an interactive picker writes .entire/settings.json with the skills and per-agent prompts to apply.

When the run starts, the command prints a scope banner, for example reviewing feat/X vs main: 3 commits, 7 files changed, 2 uncommitted, and threads the checkpoint history of every commit in scope into the agent's prompt. That gives the agent the context behind the changes before it starts reviewing.

You can also configure more than one agent, such as Claude Code and Codex, and entire review will dispatch them concurrently against the same scope, side by side in a live TUI dashboard. The review is recorded into the next checkpoint, so the metadata travels with the commit. --edit re-opens the picker; --agent NAME overrides the default.

Try Experimental Workflows

entire labs is where to find experimental Entire workflows. Run it with no arguments and you get a short list of commands we are actively refining, alongside a reminder that defaults and flag names may change.

Today the list contains a single entry: entire review, the command above. As we ship more experimental workflows, they will appear here while we refine them in the open.

Get Started

Update the CLI to 0.6.1+, then run the following commands from any enabled repo:

$entire recap
entire labs
entire review --help