TACLOR-AOS

TACLOR AOS is an autonomous software development operating system that orchestrates goals, tasks, and agents to build software.

🚀 Quickstart

1. Install dependencies:

   pip install -e .

2. Start the runtime engine:

   aos run

3. Submit a goal:

   aos goal submit <goal-id> --description "..."

4. View status:

   aos status snapshot

⚠️ The CLI uses local config files under configs/ and stores runtime state under the data root defined in configs/memory.yaml.

⚡ Quickstart: Local Development

Use two terminals so the runtime loop and API server run side by side.

1) Install dependencies

pip install -e .

2) Start the runtime engine (Terminal A)

aos run

3) Start the API server (Terminal B)

aos api --host 127.0.0.1 --port 8000

4) Verify API connectivity

curl http://127.0.0.1:8000/api/health
curl http://127.0.0.1:8000/api/status

5) Submit and inspect work (Terminal C, optional)

aos goal submit local-dev-goal --description "Build telemetry ingestion API"
aos status snapshot
curl http://127.0.0.1:8000/api/tasks

Point external tools to the API

• Base URL: http://127.0.0.1:8000
• Health: GET /api/health
• Status: GET /api/status
• Tasks: GET /api/tasks

Examples:

• Postman/Insomnia: create a request with method GET and URL http://127.0.0.1:8000/api/health
• Browser: open http://127.0.0.1:8000/api/status
• Script/client SDK: set API base URL to http://127.0.0.1:8000

🧩 Configuration

Configuration is driven from configs/runtime.yaml by default.

Profiles

You can use alternate config profiles (e.g., dev / prod) by setting:

export TACLOR_RUNTIME_CONFIG=configs/dev.yaml
# or
set TACLOR_RUNTIME_CONFIG=configs\prod.yaml

Environment overrides

Any runtime setting can be overridden via environment variables:

• DB_URL (e.g., sqlite:./data/taclor.db)
• TACLOR_RUNTIME_MODE
• TACLOR_LOOP_INTERVAL_SECONDS
• TACLOR_MAX_PARALLEL_AGENTS
• TACLOR_RETRY_FAILED_TASKS
• TACLOR_SYSTEM_INTEL_ENABLED
• TACLOR_SYSTEM_INTEL_INTERVAL

🧪 Running tests

Execute the full test suite:

python -m pytest -q

🏛️ Architecture

For a full architect-level system design, see:

• ARCHITECTURE.md

🚢 Docker / Deployment

TACLOR-AOS can run as a container via the included Dockerfile and docker-compose.yml.

Build and run (Docker)

docker build -t taclor-aos .
docker run --rm -it \
  -v "$PWD/data":/data \
  -p 8000:8000 \
  taclor-aos aos run

Run (docker-compose)

docker compose up --build

⚠️ The container mounts ./data so the SQLite runtime state persists between runs.

🧭 Runtime API Endpoints

The runtime exposes a simple HTTP API server (default: 127.0.0.1:8000). Start it with:

aos api

Common endpoints

• GET /api/health — runtime health check (OpenClaw + DB + orchestrator)
• GET /api/status — orchestrator snapshot (health, components, metrics)
• GET /api/tasks — current task list
• GET /api/agents — agent status snapshot
• GET /api/workspaces — active workspaces
• GET /api/logs — recent event log entries
• GET /api/metrics — runtime metrics snapshot
• GET /api/escalations — blocked/escalated tasks
• GET /api/git — git branch/commit info
• GET /api/validation — validation snapshot (placeholder)
• GET /api/openclaw/agents — OpenClaw agent list
• GET /api/openclaw/tasks — OpenClaw task list
• GET /api/system/performance — system performance metrics
• GET /api/system/failures — recent failure report
• GET /api/system/optimizations — recent optimization actions

🛠️ VS Code Extension (Operator Panel)

A lightweight VS Code extension is included under vscode-extension/ to interact with the AOS runtime via the CLI.

Usage

1. Open the workspace in VS Code.
2. Install and run the vscode-extension (see its README for details).