Rapid, High-Yield, Token-based, Human-in-loop, Management
A project management methodology specifically designed for AI agents with strategic human integration.
RHYTHM Method is a project management methodology that bridges the gap between traditional human-focused methodologies (Agile, Scrum, Kanban) and pure agentic processes. Unlike traditional methodologies that adapt human processes for agents, RHYTHM starts with agent capabilities and integrates human oversight strategically.
Just as a conductor (human) guides an orchestra (agents) to create harmonious music, RHYTHM Method ensures that agents work together in a coordinated, controlled flow—even when operating at fast TEMPO.
- TEMPO: The speed/pace at which agents operate (fast computational speeds)
- RHYTHM: The control, flow, and coordination that ensures quality and alignment
Key Philosophy: Agents work at a fast tempo, but RHYTHM ensures control, flow, and Human-in-the-Loop (HITL) integration.
- TEMPO: Agents operate at fast computational speeds (hours, not weeks)
- Flow: Continuous execution with dependency-driven prioritization
- Control: Human-in-the-Loop checkpoints at critical decision points
- Coordination: Multi-agent collaboration with structured interfaces
- Precision: Token-based estimation replaces abstract story points
- Adaptive: Methodology evolves based on real-time learnings
- Understand the Fundamentals: Read Core Concepts to understand RHYTHM Method
- Learn the Cycles: Review Flow Cycles to understand the execution model
- Set Up Your Project: Follow Getting Started to adopt RHYTHM Method
- Learn the Methodology: Start with Core Concepts to understand RHYTHM Method fundamentals
- Understand TEMPO: Review TEMPO Configuration to control agent speed
- Configure Your Project: Use Project Setup to create project configuration
Start here to understand RHYTHM Method fundamentals:
- Core Concepts - Agents (BA/RA/WA), TEMPO, HITL, Token estimation
Learn how RHYTHM Method orchestrates work:
- Flow Cycles - The 6 main cycles (Project Init, Feature Spec, Work Unit Creation, Challenge, Task Breakdown, Task Execution)
- Special Cycles - Quality Check and Dependency-driven Prioritisation cycles
- Processes - Supporting processes (Project Interview, Feature Interview, Cycle Review, HITL Check)
- State Management - State transitions for Projects, Features, Work Units, Agent Tasks, Bugs
Understand how work is structured and estimated:
- Work Breakdown Structure - Project, Feature, Work Unit, Agent Task, Bug hierarchy
- Token Estimation - Token-based estimation methodology and formulas
- Dependency Management - Dependency-driven prioritization and detection
Configure RHYTHM Method for your project:
- TEMPO Configuration - Understanding and configuring TEMPO levels
- Project Setup - Project.Manifest and Project.Config templates and setup
Best practices and real-world application:
- Getting Started - Step-by-step guide to adopting RHYTHM Method
- Best Practices - Proven patterns and recommendations
- Common Challenges - Common challenges and solutions
- Error Handling - Error handling and failure recovery strategies
Quick reference and examples:
- Workflow Examples - End-to-end workflow walkthroughs
- Quick Reference - Formulas, decision trees, and key information
- Dictionary - Comprehensive terminology guide
- Speed: Execution cycles (hours) instead of sprints (weeks)
- Precision: Token-based estimation instead of abstract story points
- Automation: Automated dependency analysis and prioritization
- Structure: Machine-readable specifications and workflows
- Efficiency: No unnecessary meetings or delays
- Human-in-the-Loop (HITL) Integration: Human oversight at critical decision points
- Strategic Input: Human judgment for business decisions
- Context: Human-readable explanations and rationale
- Control: Configurable approval gates and overrides
- Collaboration: Human-AI partnership, not replacement
Traditional project management methodologies were designed for human teams. When applied to AI agent teams, they create friction, overhead, and missed opportunities. Conversely, pure agentic processes designed for computational efficiency exclude humans and create information overload.
RHYTHM Method bridges this gap by creating a methodology that leverages agent capabilities while maintaining essential human oversight and strategic input.
RHYTHM Method is best suited for:
- Agentic development environments with AI agents capable of development work
- Projects requiring fast iteration and rapid delivery cycles
- Teams comfortable with automation and structured workflows
- Projects with clear dependencies that can be systematically managed
- Organizations willing to invest in training and methodology adoption
RHYTHM Method may not be ideal for:
- Traditional human-only teams without AI agent infrastructure
- Projects requiring extensive human collaboration at every step
- Teams resistant to change or new methodologies
- Projects with unclear or highly ambiguous requirements
- Organizations without agent infrastructure or AI development capabilities
Recommendation: Start with a pilot project to evaluate RHYTHM Method's fit for your organization. Use Moderate TEMPO initially, focus on well-defined work, and build historical data to improve estimation accuracy over time.
RHYTHM Method is an alternative to:
- Spec-Driven Development (SDD)
- Scrum
- Agile
- Kanban
- Traditional project management
RHYTHM Method is not a modification of these methodologies—it's a new methodology designed specifically for agents with human integration.
This repository contains pure documentation in a numbered format for easy sequential reading:
RHYTHM-Method/
├── 01-core-concepts.md # Fundamental concepts (Agents, TEMPO, HITL, Tokens)
├── 02-flow-cycles.md # Main execution cycles
├── 03-special-cycles.md # Quality Check and Dependency-driven Prioritisation
├── 04-processes.md # Supporting processes
├── 05-state-management.md # State transitions for all work items
├── 06-work-breakdown-structure.md # Work hierarchy (Project → Feature → Work Unit → Task)
├── 07-token-estimation.md # Token-based estimation methodology
├── 08-dependency-management.md # Dependency detection and prioritization
├── 09-tempo-configuration.md # TEMPO levels and configuration
├── 10-project-setup.md # Project.Manifest and Project.Config setup
├── 11-getting-started.md # Adoption guide
├── 12-best-practices.md # Proven patterns and recommendations
├── 13-common-challenges.md # Common issues and solutions
├── 14-error-handling.md # Error handling and recovery
├── 15-workflow-examples.md # End-to-end workflow examples
├── 16-quick-reference.md # Quick reference guide
├── 17-dictionary.md # Comprehensive terminology guide
├── CONTRIBUTING.md # Contribution guidelines
├── README.md # This file
└── LICENSE # MIT License
Reading Path:
- Quick Start: 01 → 02 → 11 (Core Concepts, Flow Cycles, Getting Started)
- Complete Understanding: Read sequentially 01-17
- Implementation Focus: 10 → 11 → 12 → 15 (Setup, Getting Started, Best Practices, Examples)
- Reference: 16 for quick lookups, 17 for terminology
Contributions are welcome! Please see our Contributing Guidelines for details.
We especially welcome feedback from implementers! RHYTHM Method is a first-of-its-kind methodology, and your real-world implementation experience is invaluable for improving the methodology. Whether you're using Baton Framework or adapting RHYTHM Method to another system, we want to hear about:
- What works well in practice
- What doesn't work or needs clarification
- Edge cases or scenarios not covered
- Suggestions for improvements
- Implementation challenges and how you solved them
Please open an issue to share your feedback. Your input helps make RHYTHM Method better for everyone.
RHYTHM Method was originally designed for the Baton Framework, which provides:
- Foundational Architecture: The instruction architecture that underlies RHYTHM Method
- Tooling: Automated workflows, message formats, and integration capabilities
- Platform: Full Agentic Development Environment with RHYTHM Method workflows built-in
While RHYTHM Method works best with Baton Framework, the methodology itself is framework-agnostic and can be adapted to other agent orchestration systems. The documentation and principles are designed to be implementation-agnostic.
- Baton Framework - The foundational instruction architecture for RHYTHM Method
- Baton CLI - Tooling for scaffolding RHYTHM Method into projects
- Baton Platform - Full Agentic Development Environment (Coming Soon)
This project is licensed under the MIT License - see the LICENSE file for details.
RHYTHM Method is developed by Conductus Labs Ltd, a UK-based AI application layer company.
- Website: conductuslabs.com
- Product: Baton Platform - AI Orchestration Platform (SaaS) - Coming Soon
- GitHub: github.com/Conductus-Labs
Built with ❤️ in the UK