augint-tools 4.2.1

CLI for AI-assisted repository and workspace workflows

augint-tools

CLI orchestration layer for AI-assisted repository and workspace workflows.

augint-tools provides a stable, machine-parseable command surface for humans and AI agents to coordinate development workflows across single repositories and multi-repo workspaces. It is designed to be called directly by AI skills, replacing ad-hoc shell scripts with reliable, JSON-enabled commands.

Features

• Dual-mode operation: Commands for both single repos (ai-tools repo) and workspaces (ai-tools workspace)
• AI-first design: Every command supports --json output for agent parsing
• Repo-type aware: Understands library, service, and workspace repository patterns
• Safe defaults: No destructive git operations without explicit commands
• GitHub integration: Issue management, PR creation, CI status monitoring

Installation

pip install augint-tools

Or with uv:

uv tool install augint-tools

Quick Start

Single Repository Workflows

# Initialize repo metadata
ai-tools init --library

# Check repository status
ai-tools repo status --json

# Search/pick issues
ai-tools repo issues pick "bug"

# Create feature branch
ai-tools repo branch prepare --issue 42 --description "fix the thing"

# Run checks
ai-tools repo check run
ai-tools repo check run --preset full --fix

# Submit work (push + create PR)
ai-tools repo submit

Workspace Workflows

# Initialize workspace
ai-tools init --workspace

# Sync all child repositories
ai-tools workspace sync --json

# Check status across all repos
ai-tools workspace status

# Create coordinated branches
ai-tools workspace branch --name feat/multi-repo-change

# Run tests and lint across all repos
ai-tools workspace test
ai-tools workspace lint

# Run command in all repos
ai-tools workspace foreach -- git status

# Submit PRs for all modified repos
ai-tools workspace submit

Command Reference

Top-Level Commands

• ai-tools init [--library|--service|--workspace] - Initialize repository metadata

Repository Commands (ai-tools repo)

• inspect - One-call repo snapshot (kind, branch, toolchain, command plan)
• status - Show repository status (branch, dirty state, PRs, CI)
• issues pick [query] - Issue recommendation and search
• branch prepare - Create work branch from correct base
• check plan - Resolve validation plan without running
• check run - Execute validation plan
• submit - Push branch and create PR with automerge
• ci watch - Monitor CI run
• ci triage - Classify CI failures

Workspace Commands (ai-tools workspace)

• inspect - Workspace snapshot
• status - Status across all child repositories
• sync - Clone missing repos and update existing
• issues [query] - Aggregate issues from all repos
• branch - Create coordinated branches
• check - Grouped validation across repos
• test - Alias for check --phase tests
• lint - Alias for check --phase quality
• foreach <command> - Execute command in all repos
• submit - Push and create PRs for modified repos

Configuration

Repository Classification

ai-shell.toml:

[project]
repo_type = "library"        # or "service", "workspace"
branch_strategy = "main"     # or "dev"
dev_branch = "dev"           # when branch_strategy = "dev"

Workspace Manifest

workspace.yaml (for workspace repos):

workspace:
  name: my-workspace
  repos_dir: repos

repos:
  - name: my-lib
    path: repos/my-lib
    url: https://github.com/org/my-lib.git
    repo_type: library
    base_branch: main
    pr_target_branch: main
    install: uv sync --all-extras
    test: uv run pytest -v
    lint: uv run pre-commit run --all-files
    depends_on: []

Development

Setup

uv sync --all-extras

Running Tests

uv run pytest                    # Run all tests
uv run pytest --cov             # With coverage
uv run pytest -k test_name      # Specific test

Code Quality

uv run ruff check src/ tests/   # Lint
uv run ruff format src/ tests/  # Format
uv run mypy src/               # Type check
uv run pre-commit run --all-files  # All hooks

Design Principles

1. Human and AI first - Commands work well for both interactive use and programmatic calls
2. JSON always available - Every orchestration command supports stable --json output
3. Safe defaults - No destructive behavior without explicit confirmation
4. Repo-type aware - Different defaults for libraries, services, and workspaces
5. Skills call tools - AI skills orchestrate this CLI, not replace it with shell scripts

License

MIT