Autobots DevTools Shared Library

Dynamic Agent (Dynagent) is the core of this library. It turns your prompts and business processes into production-ready, multi-agent applications—chatbots and unsupervised workflows—in hours. You focus on prompts, output schemas, and domain logic; Dynagent handles multi-LLM wiring, UI integration, observability, and batch processing out of the box.

Quickstart Build your first agent by following Scaffolding guide here.

Essential features

Batteries included

Quickstart

How-to guides

Advanced

Workspace setup

This library is part of a multi-repository workspace. Use a shared virtual environment at the workspace root.

Prerequisites: Python 3.12+, Poetry (e.g. brew install poetry on macOS).

1. Clone the workspace

cd /path/to/your/work
git clone <workspace-url> ws-jarvis
cd ws-jarvis

2. Create shared virtual environment

make setup

This creates a shared .venv at the workspace root.

3. Clone this repository

git clone https://github.com/Pratishthan/autobots-devtools-shared-lib.git
cd autobots-devtools-shared-lib

4. Install dependencies

make install-dev   # with dev dependencies (recommended)
# or
make install       # runtime only

5. Pre-commit hooks

make install-hooks

Development

Run from autobots-devtools-shared-lib/:

# Testing
make test              # with coverage
make test-fast         # no coverage
make test-one TEST=tests/unit/test_example.py::test_function

# Code quality
make format            # Ruff format
make lint              # Ruff lint (auto-fix)
make check-format      # check only
make type-check        # Pyright
make all-checks        # format, type, test

# Dependencies & build
make install / make install-dev / make update-deps
make build
make clean
make help

Workspace-level commands

From the workspace root:

make test
make lint
make format
make type-check
make all-checks

Project structure

autobots-devtools-shared-lib/
├── src/autobots_devtools_shared_lib/
│   ├── dynagent/           # Multi-agent framework
│   ├── chainlit_ui/        # Chainlit UI components
│   ├── llm_tools/          # LLM integrations
│   ├── observability/      # Observability helpers
│   └── batch_processing/   # Batch utilities
├── tests/
│   ├── unit/
│   ├── integration/
│   └── e2e/
├── .github/workflows/
├── pyproject.toml
├── poetry.toml             # Poetry settings (uses workspace .venv)
├── Makefile
├── CONTRIBUTING.md
└── PUBLISHING.md

Code quality standards

• Type safety: Type annotations; Pyright (basic mode).
• Testing: pytest; unit, integration, e2e.
• Formatting: Ruff, line length 100.
• Linting: Ruff, strict rules.
• Pre-commit: Format, lint, type-check, tests on commit.

Testing

make test
make test-one TEST=tests/unit/test_example.py
make test-cov   # HTML coverage report

• Unit (tests/unit/): Functions and classes.
• Integration (tests/integration/): Component interactions.
• E2E (tests/e2e/): Full workflows.

Type checking

make type-check

All code must have type annotations. Pyright runs in basic mode.

Contributing

See CONTRIBUTING.md for development workflow and guidelines.

Publishing

See PUBLISHING.md for PyPI publishing.

License

MIT

Authors

• Pra1had — GitHub · pralhad.kamath@pratishthanventures.com

Questions?

Open an issue on the project repository.