Agentic Maya
Agentic Maya — Governed multi-agent orchestration with YAML pipelines, policies, and skills for enterprise-grade AI workflows.
What It Does
Agentic Maya provides a secure, auditable framework for running multi-agent AI pipelines with built-in governance, memory management, and human oversight. It's designed for production use cases requiring compliance, traceability, and safety.
Key Features
- YAML-Driven Pipelines: Define complex workflows with steps, checkpoints, and fallback models
- Policy Enforcement: Six-layer governance system (Intent, Auth, Budget, PII, External Gate, Escalation)
- Skill Registry: Reusable, versioned agent capabilities with audit trails
- Memory Tiers: Session, episodic, long-term, and audit memory for safe persistence
- CLI Tools: Full lifecycle management with status, audit, checkpoint, and rollback commands
- HITL Integration: Human-in-the-loop approvals for critical decisions
- Audit Logging: Immutable, hash-chained logs for compliance
Quick Start
Installation
pip install agentic-maya
# or for development
pip install -e .
First Pipeline
# Run a sample pipeline
maya run ./pipelines/demo_pipeline.yml
# Check execution status
maya status <session_id>
# View detailed audit trail
maya audit <session_id>
CLI Commands
maya run <pipeline.yml>- Execute a pipelinemaya status <session_id>- Check pipeline statusmaya audit <session_id>- View audit logsmaya checkpoint <session_id> --label <name>- Create checkpointmaya rollback <session_id> --to <checkpoint>- Rollback to checkpointmaya approve <session_id> --request-id <id>- Approve pending requestsmaya pause <session_id>- Pause executionmaya terminate <session_id>- Stop pipeline
Architecture
Governance Layers
- Intent Verification - Validate task intent against policies
- Model Authorization - Check model access permissions
- Budget Control - Enforce token and cost limits
- PII Detection - Block sensitive data leaks
- External Gate - Control API/tool usage
- Escalation - Human approval for critical actions
Memory Tiers
- Session Memory: Current execution state
- Episodic Memory: Step checkpoints and resumes
- Long-Term Memory: Persistent knowledge and preferences
- Audit Memory: Compliance event history
Agent Roles
- Orchestrator: Controls pipeline execution
- Specialist: Executes domain tasks
- Subagent: Scoped helper agents
- Critic: Validates outputs and compliance
- Memory Manager: Handles all memory operations
- Tool Executor: Runs external tools/APIs
- Audit Agent: Enforces governance and logging
Documentation
📖 Full Documentation - Complete guides and API reference
Project Structure
agentic-maya/
├── main.py # CLI entry point
├── pyproject.toml # Project configuration
├── maya.toml # Runtime defaults
├── src/maya/ # Core package
│ ├── cli.py # Command-line interface
│ ├── runtime.py # Pipeline execution engine
│ ├── governance.py # Policy enforcement
│ ├── memory.py # Memory management
│ └── schemas.py # Data validation
├── pipelines/ # YAML pipeline definitions
├── policies/ # YAML governance policies
├── skills/ # YAML skill definitions
├── tests/ # Test suite
│ ├── test_*.py # Unit tests
│ └── scenarios/ # Integration tests
├── docs/ # Markdown documentation
└── website/ # Static website
├── index.html # Homepage
└── docs/ # HTML documentation
Development
Setup
git clone <repository-url>
cd agentic-maya
pip install -e .[dev]
Testing
# Run all tests
python -m pytest
# Run specific test
python -m pytest tests/test_governance.py -v
# Run with coverage
python -m pytest --cov=src/maya
Building Documentation
# Generate HTML docs from Markdown
python scripts/build_docs.py
Contributing
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Write tests for new functionality
- Ensure all tests pass (
python -m pytest) - Commit changes (
git commit -m 'Add amazing feature') - Push to branch (
git push origin feature/amazing-feature) - Open a Pull Request
Guidelines
- Follow PEP 8 style guidelines
- Add type hints for new functions
- Update documentation for API changes
- Ensure backward compatibility
License
This project is licensed under the MIT License - see the LICENSE file for details.
Support
- 📧 Email: adlitoxit@gmail.com
- 🐛 GitHub: GitHub
Agentic Maya - Safe, Auditable, Multi-Agent AI Orchestration
Metadata
Release files for agentic-maya 0.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| agentic_maya-0.2.0.tar.gz | 32.0 kB | Without provenance | Details for agentic_maya-0.2.0.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| agentic_maya-0.2.0-py3-none-any.whl | Python 3 | none | any | Details for agentic_maya-0.2.0-py3-none-any.whl |
Total release size: 65.7 kB
Release files / agentic_maya-0.2.0.tar.gz
| Download URL | agentic_maya-0.2.0.tar.gz |
|---|---|
| Size | 32.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
0bc823e13f6a426bd110ad2a5254005bf44578910846db4b30425775c7e42abf
|
|
BLAKE2b-256 checksum How to use checksums |
c3484ea62799fcbcdb498ac36cabed147449cffcec07592dbba3eef9562f2b7c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.10
|
Release files / agentic_maya-0.2.0-py3-none-any.whl
| Download URL | agentic_maya-0.2.0-py3-none-any.whl |
|---|---|
| Size | 33.7 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
faeb1a25d27081c7081089860610651c693f305c6872b336e14cd56bebb18176
|
|
BLAKE2b-256 checksum How to use checksums |
3824d2898551dd150eaf4eb5af3a9c48c8a5f546051dbdb4ca47f59b6ae4bfae
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.12.10
|
Security for agentic-maya 0.2.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.