Workflow
The Hydro methodology operates with a single authoritative workflow state field that serves as the source of truth for all workflow logic. This ensures consistency and eliminates parallel tracking systems that can create confusion.
Standard Workflow States
The methodology defines seven primary states that tasks flow through during their lifecycle
Backlog
Initial capture of requirements and ideas
Basic description exists, dependencies undefined
Business priority for analysis
Duration: indefinite until prioritized for refinement
In Refinement
Define all needed info for development
AI classification, context package, acceptance criteria
Human validates classification, AI suggests based on complexity
Complete when all specs done, dependencies mapped
Ready for Dev
All prerequisites satisfied, ready for implementation
Clear acceptance criteria, AI context prepared, no blocking dependencies
All dependency tasks completed, specifications validated
Can start when developer/AI capacity available, ranked by unlock potential
Blocked
Temporary state for tasks that cannot proceed
External blockers identified and documented
Regular review to ensure blockers being addressed
Return to appropriate state when blockers resolved
In Progress
Active development work occurring
Only tasks within currently active wave, typically one task at once
Wave becomes active when first task transitions
AI-ready execute autonomously, others require human collaboration
In Review
Completed implementation awaiting human validation
Code quality, business logic accuracy, integration compatibility review
Automated tests, security scans, performance validation
Review depth varies by AI classification, human sign-off required
Done
Completed, tested, and validated work ready for production
All acceptance criteria met, quality gates passed
Successfully integrated, documentation updated
Triggers analysis of dependent tasks for potential unblocking
State Transition Rules
Each state change requires specific conditions to be met, preventing premature advancement and ensuring quality standards. The methodology enforces these validation gates to maintain workflow integrity and prevent coordination failures.
Backlog → In Refinement:
Task selected for analysis based on business priority
Basic requirements and context available
Team capacity available for refinement work
No blocking dependencies at refinement level
In Refinement → Ready for Dev:
AI classification assigned and validated by human reviewer
Dependencies clearly mapped and documented in proper format
Acceptance criteria complete and testable
AI context package prepared (for AI-classified tasks)
Epic assignment confirmed and validated
Wave assignment calculated or manually set
Ready for Dev → In Progress:
All dependency tasks must be in "Done" status (strict enforcement)
Team member or AI capacity available for task execution
Task must be within currently active wave (Active Wave Singularity enforcement)
No higher priority tasks available in ready state within the same wave
Human approval for task initiation (where required by classification)
In Progress → In Review:
Implementation complete according to all acceptance criteria
Automated tests passing with required coverage thresholds
Code committed to version control and available for review
No blocking issues or unresolved technical debt introduced
AI tasks may auto-transition with automated validation
In Review → Done:
Human validation complete for quality and business logic accuracy
Integration tests passing at both task and wave level
Documentation updated and accurate
Performance and security requirements met and validated
All Definition of Done criteria satisfied
No breaking changes to existing interfaces (or properly managed if necessary)
In Review → In Progress (rework scenario):
Validation identifies specific issues requiring correction
Detailed feedback provided for rework with clear requirements
Human reviewer approval for returning to development
Rework scope defined and estimated
Ready for Dev/In Progress → Blocked:
External dependencies or blockers clearly identified and documented
Blocking issues have defined resolution approach and ownership
Impact assessment completed for timeline and dependent tasks
Regular review schedule established for blocker resolution
Blocked → Ready for Dev/In Progress:
All blocking issues resolved and validated
Dependencies satisfied or acceptable workarounds implemented
Human approval for unblocking and continuation
Task returns to appropriate state based on completion status when blocked
Quality Gates by State:
In Refinement: Classification accuracy, dependency completeness, wave assignment validation
Ready for Dev: Prerequisite validation, wave boundary compliance
In Progress: Active wave enforcement, resource allocation validation
In Review: Code quality, business logic accuracy, integration compatibility
Done: Complete validation against Definition of Done criteria
Note: advanced implementations may include automated cascade analysis for dependency unblocking and priority calculation for optimal task sequencing, empowering the methodology through specialized tooling.
Last updated