gre

A super fast multi repo git recap for AI agents and speed obsessed humans.

Run one command:

gre

No prompts. No interactive mode. No extra setup after config.

Install

cargo install gre
# or
brew install tappunk/gre/gre

CLI

gre [--config PATH] [--json] [init [--force]]

Minimal contract:

• gre prints the default human recap.
• gre --json prints machine-readable output for agents/scripts.
• gre init writes ~/.config/gre/config.toml.
• gre init --force overwrites an existing config.
• gre --config PATH uses an explicit config path.

Config

Default path:

~/.config/gre/config.toml

Example config:

paths = [
  "~/src/gre",
  "~/src/muthr",
]

[output]
default_json = false

Alternative repo shapes are also supported:

repos = ["~/src/gre", "~/src/muthr"]

[[repos]]
name = "gre"
path = "~/src/gre"

Human Output

gre sorts repos by action priority: conflicts -> dirty -> behind -> ahead -> clean.

ahead/behind is local-tracking based. gre does not fetch remotes. Run git fetch in a repo to refresh tracking refs before running gre.

The summary line includes run timing (time, avg) to track command speed.

Columns:

repo  branch  sync  state  next  last_commit  path

next values:

• resolve-conflicts
• commit-or-stash
• pull
• push
• sync
• none

JSON Output

gre --json is stable for automation and includes:

• schema_version
• summary counts (configured_total, succeeded_total, failed_total, total)
• summary repo state (dirty, behind, ahead)
• summary timing (elapsed_ms, avg_repo_ms)
• summary tail latency (p95_repo_ms)
• repos array

Example:

{
  "schema_version": "1",
  "summary": {
    "total": 2,
    "dirty": 1,
    "behind": 0,
    "ahead": 1
  },
  "repos": [
    {
      "name": "gre",
      "path": "/Users/alice/src/gre",
      "branch": "main",
      "ahead": 1,
      "behind": 0,
      "staged": 0,
      "unstaged": 2,
      "untracked": 0,
      "conflicts": 0,
      "clean": false,
      "last_hash": "abcd123",
      "last_subject": "tune output",
      "last_relative": "2 hours ago"
    }
  ]
}