Lifecycle

Overview

Every project in Romyq follows the same eight-stage lifecycle. These stages are ordered — each one depends on the previous completing. The complexity profile determines how many phases appear within stage 03 and how many tasks are generated per phase.

The Eight Stages

01 · Mission

The operator defines what the project must accomplish. The mission is a plain-language statement stored in mission.md at the project root.

Build a REST API for analytics with authentication and PostgreSQL.
Use Hono on Cloudflare Workers. All endpoints must have integration tests.

The mission is passed to the planning provider in every prompt. Planning decisions that contradict the mission are rejected.

02 · Lifecycle

Romyq selects a lifecycle model based on the project type and the complexity profile chosen during romyq init. The lifecycle model determines the shape of the stages that follow.

03 · Phases

The planning provider decomposes the mission into ordered phases. Each phase has:

• A name and description
• Done criteria (what must be true before the phase is complete)
• A dependency on all predecessor phases

The number of phases depends on the complexity profile:

Phases cannot overlap. Phase N cannot start until Phase N−1 is complete and its done criteria are satisfied.

04 · Tasks

Within each phase, the planning provider generates tasks. Each task has:

• A title and description
• A capability tag (what the project gains when the task is done)
• Dependencies on other tasks within the phase
• A reference to which phase it belongs

Tasks pass through governance evaluation before they enter the execution queue. Tasks that violate the project rules are rejected and replanned.

05 · Execution

The execution provider (Claude Code) runs each task. After every task completes, Romyq:

1. Records the outcome in .romyq/history.json
2. Updates the knowledge layer
3. Recomputes the readiness score
4. Emits a recommendation (Continue, Pause, Review, or Stop)

The operator observes this in the dashboard. Execution does not stop automatically.

06 · Validation

After the execution phase tasks are complete, Romyq checks the done criteria for each phase. This is not a test runner — it evaluates whether the criteria defined during planning are satisfied based on the current project state.

If a phase fails validation, the execution provider is prompted to close the gap before moving on.

07 · Readiness

Romyq computes a readiness score from 0 to 100 after every task, not just after validation. The score is a weighted aggregate across four categories:

A score of 100 means all criteria are met, all required capabilities are present, and there are no active violations.

08 · Completion

The recommendation engine emits Stop when:

• All phases are complete
• All done criteria are satisfied
• Readiness is ≥ 75

The Stop recommendation is advisory. The operator confirms by running romyq stop. The lifecycle is recorded as complete in .romyq/history.json.

Viewing lifecycle state

romyq roadmap              # show phase status and progress
romyq readiness            # show current readiness score
romyq readiness --json     # machine-readable output
romyq timeline             # chronological log of all lifecycle events

Resuming after interruption

Lifecycle state is preserved on failure or interruption. Resume with:

romyq resume

Romyq reads .romyq/state.json to determine where to restart from. Previously completed tasks are not re-run.