a3 0.2.2

Foundational scaffolding for the A4 AI workflow orchestrator

A4

Foundational scaffolding for the A4 AI workflow orchestrator. This repository hosts the CLI and SDK skeleton along with shared infrastructure (configuration, dependency injection, logging) that other phases will extend.

Getting Started

1. Create a virtual environment with Python 3.10+.
2. Install the package in editable mode:

   pip install -e .[dev]
   

3. Run diagnostics:

   a4 --health
   a4 --version
   

Highlights

• Provider Fallbacks: Configure multiple AI providers with automatic failover in a4.yaml (ai.fallback_providers).
• Resumable Drafting: Retry interrupted drafting runs with a4 draft generate --resume; progress is checkpointed under .a4/state/.
• Usage Analytics: Telemetry-aware logging writes to .a4/metrics/. Inspect summaries with a4 observe usage or disable tracking via features.telemetry=false.

Project Structure

• a4/ - Python package with CLI, configuration loader, and core utilities.
• .a4/ - Workspace artifacts generated at runtime (plans, specs, drafts, validation).
• docs/ - Architectural, product, and delivery documentation.
• tests/ - Automated tests covering core scaffolding.

See docs/ for detailed design and roadmap information.

Documentation

• docs/CLI_HANDBOOK.md - CLI usage guide with full command reference.
• docs/SDK_QUICKSTART.md - Programmatic examples for integrating the SDK.
• docs/ACCEPTANCE_REPORT.md - Pilot testing summary and sign-off details.
• docs/POST_LAUNCH_SUPPORT.md - Support rotation and telemetry hand-off plan.
• docs/SYSTEM_ARCHITECTURE.md - Architecture blueprint with release and operations view.