Festival Methodology Overview

Festival Methodology is a collaborative, step-oriented planning system built for human-AI development workflows. It replaces traditional project management constructs -- sprints, story points, velocity tracking, time estimates -- with a model that thinks in steps to goals. Every unit of work is defined by what must be accomplished and what completion looks like, not by how long someone guesses it will take.

AI-human collaboration operates at a fundamentally different speed than traditional teams. When an AI agent executes implementation steps at 30x-100x the pace of manual development, duration-based planning becomes noise. Festival Methodology strips that noise out. What remains is a clean focus on logical steps, dependencies between those steps, opportunities for parallel execution, and unambiguous completion criteria.

The result is a planning system that works for a solo developer driving an AI agent through a complex build, a team coordinating multi-phase projects, or an AI agent operating autonomously across long-running sessions.

Core Principles

1. Goal-Oriented Step Planning -- Every festival exists to achieve a specific, measurable goal. Planning starts by identifying the steps required to reach that goal, not by estimating calendars.
2. Human-AI Collaborative Planning -- Humans provide goals, requirements, architectural decisions, and validation. AI agents identify steps, structure them, create task specifications, and execute autonomously.
3. Requirements-Driven Implementation -- Implementation sequences can only be created after requirements are defined. No guessing. No premature coding.
4. Just-in-Time Sequence Creation -- Sequences are designed when a phase is ready for execution, not months in advance. Plans stay grounded in current understanding.
5. Hyper-Efficient AI Execution -- The methodology feeds work to AI agents in structured, unambiguous units that maximize autonomous execution at unprecedented speed.
6. Step-Based Progression -- Progress is measured by completed steps with verified outcomes, not by hours logged or percentage-of-time-elapsed.
7. Context Preservation -- Documentation is read just-in-time. Festival structure preserves context across sessions so agents resume work without re-reading the entire project history.
8. Quality Gates -- Every implementation sequence ends with mandatory quality verification steps (testing, review, iteration). Quality is structural, not optional.
9. Extensible Methodology -- The core hierarchy is fixed, but phase types, festival types, quality gate configurations, and workflow patterns are all extensible.

Step-Based vs Time-Based Thinking

Traditional PM and Festival Methodology solve the same problem but start from opposite assumptions.

Traditional PM focuses on: time estimates and schedules, duration-based sprint planning, resource allocation over calendar periods, velocity and burndown charts.

Festival Methodology focuses on: logical steps toward goal achievement, completion criteria for each step, dependencies between steps, parallel execution opportunities.

When planning work, the key questions are:

• What steps are needed? -- Identify every step required to reach the goal.
• What is the logical order? -- Determine which steps depend on others.
• What can be done in parallel? -- Steps with the same numeric prefix execute concurrently.
• How do we know each step is complete? -- Define verification criteria, not deadlines.

"How long will this take?" is replaced by "What steps will achieve this goal?" -- a question that produces actionable structure instead of speculative estimates.

Human-AI Collaborative Workflow

Festival Methodology is not about AI pre-planning entire projects in isolation. It is a collaborative model with clear responsibilities.

Human provides: goals and success criteria, requirements and specifications, architectural decisions and technology choices, feedback and iteration guidance, scope decisions and priority calls.

AI agent provides: step identification from requirements, logical structure and sequencing, detailed task specifications with completion criteria, autonomous execution of implementation steps, progress documentation and tracking.

Humans define what needs to be achieved. AI agents structure how to get there and execute the work. Humans validate results and provide course corrections. Neither side works in isolation -- the methodology enforces this interaction through its structure.

The Three-Level Hierarchy

Festival
└── Phase       (3-digit prefix: 001_, 002_, 003_)
    └── Sequence    (2-digit prefix: 01_, 02_, 03_)
        └── Task        (2-digit prefix: 01_, 02_, 03_)

Phases are major stages of work representing significant milestones -- planning, implementation, review. Each phase has a type that determines its internal structure. Phases execute sequentially unless they share the same numeric prefix, which signals parallel execution.

Sequences are ordered groups of related tasks within a phase. A sequence represents a cohesive unit of work -- "build the authentication system" or "define the API contracts." Implementation sequences always end with quality gate tasks.

Tasks are the atomic units of work. Each task is a single markdown file with clear instructions, acceptance criteria, and a concrete deliverable. Tasks are what AI agents execute.

See: Phases | Sequences | Tasks

Festival Types

fest create festival --type standard "my-project"
fest types festival                    # discover all types
fest types festival show <type>        # see type details

Festival Lifecycle

Festivals move through status directories representing their current state. Movement is explicit -- a festival is physically moved from one directory to another.

planning/  -->  ready/  -->  active/  -->  dungeon/completed/
                                     -->  dungeon/archived/
                                     -->  dungeon/someday/

When to Use Festival Methodology

Festival Methodology is the right fit when:

• You have a clear goal and want structured steps to get there.
• You want to minimize process overhead -- no standups, sprints, or retros unless you explicitly add them.
• The work has natural dependencies where some steps must complete before others begin.
• You need flexible scope -- the ability to add, remove, or reorder steps as understanding evolves.
• Team members (human or AI) can work independently on well-defined units.
• You are working solo with AI agents or collaboratively across a team.

Key Advantages

No process overhead. No mandatory ceremonies. No standups, no sprint planning, no retrospectives -- unless you decide they add value and create a ritual festival for them.

Clear dependencies. Sequential directory numbering makes dependencies obvious. 001_PLAN completes before 002_IMPLEMENT. 01_foundation completes before 02_features. The filesystem is the dependency graph.

Flexible scope. Sequences are created just-in-time. If requirements change, add new sequences, reorder existing ones, or remove work that is no longer needed. The structure adapts to the project.

Goal achievement focus. Every element in the hierarchy has a goal document defining what success looks like. Work is complete when the goal is met, not when a calendar date arrives.

Deep understanding required upfront. The methodology front-loads thinking. Requirements must be defined before implementation sequences are created. The investment in understanding pays off in fewer false starts and less rework.