First 30 Minutes with GAIA — Full Lifecycle End-to-End
This tutorial walks the entire GAIA lifecycle from a fresh project to a closed sprint: bootstrap → discovery → planning → solutioning → first sprint (implement, review, transition, close) → planning the next sprint. The first ten steps get you to your first story being implemented; Steps 11-18 take that story through reviews, mark it done, run the sprint review and retro, close the sprint, and plan the next one.
This walkthrough takes you from a blank directory to a running sprint. Each step builds on the previous one, producing an artifact that feeds the next command. By the end, you will have a product brief, a PRD, an architecture document, a backlog of stories, a planned sprint, and your first story implemented with tests.
Have an existing codebase? Use the
First 30 Minutes (Brownfield)
companion instead -- it collapses Steps 2--7 below into a single
/gaia-brownfield invocation that reverse-engineers the
PRD, architecture, and epics from your code, then re-joins this
walkthrough at Step 8.
Overview
The walkthrough follows GAIA's five lifecycle phases: Analysis (steps 1--4), Planning (step 5), Solutioning (steps 6--7), and Implementation (steps 8--10). Deployment is covered separately in the Recipes page.
Step 1 -- Install the plugin
Install the GAIA plugin into your Claude Code environment. Refer to
the GAIA repository README for platform-specific installation
instructions. Once installed, all /gaia-* commands
become available in any Claude Code session.
Expected duration: 2 minutes.
Step 2 -- Bootstrap your project
/gaia-init
GAIA asks a series of discovery questions -- project name, tech
stack, platforms, compliance regimes, environments, and CI platform.
Your answers are saved to .gaia/config/project-config.yaml,
the single configuration file that every other GAIA command reads.
Artifact produced: .gaia/config/project-config.yaml
Expected duration: 3 minutes.
Step 3 -- Explore the problem
/gaia-brainstorm
Describe what you want to build. GAIA facilitates a structured brainstorming session, helping you explore the problem space, user needs, and potential solutions. The output captures your ideas in a format that the next commands can consume.
Artifact produced: .gaia/artifacts/creative-artifacts/brainstorm-{slug}.md
Expected duration: 5 minutes.
Step 4 -- Capture the vision
/gaia-product-brief
Synthesize the brainstorm into a structured product brief covering vision, target users, the core problem, proposed solution, scope boundaries, risks, competitive landscape, and success metrics.
Artifact produced: .gaia/artifacts/planning-artifacts/product-brief.md
Expected duration: 5 minutes.
Step 5 -- Produce requirements
/gaia-create-prd
The PM agent (Derek) transforms your product brief into a full
Product Requirements Document. The PRD contains functional
requirements (FRs), non-functional requirements (NFRs), user
journeys, data requirements, integrations, and success criteria.
Every FR gets a unique identifier (e.g., FR-001) that
flows through the rest of the lifecycle.
Artifact produced: .gaia/artifacts/planning-artifacts/prd.md
Expected duration: 5 minutes.
Step 6 -- Design the system
/gaia-create-arch
The architect agent (Theo) reads the PRD and produces an architecture document covering technology selection, system components, data architecture, API design, infrastructure, security architecture, and ADRs. Each architecture decision traces back to a PRD requirement.
Artifact produced: .gaia/artifacts/planning-artifacts/architecture.md
Expected duration: 5 minutes.
Step 7 -- Break into deliverable units
/gaia-create-epics
GAIA breaks the architecture into epics and stories. Each epic groups related stories that together deliver a significant capability. Each story gets acceptance criteria, a risk rating, and dependency links.
Artifact produced: .gaia/artifacts/planning-artifacts/epics-and-stories.md
Expected duration: 3 minutes.
Step 8 -- Prepare individual stories
/gaia-create-story E1-S1
Creates a detailed story file for a specific story from the epics-and-stories document. The file includes YAML frontmatter (status, risk, dependencies), acceptance criteria, subtasks, a Definition of Done checklist, and a Review Gate table. Repeat for each story you want to work on.
Artifact produced: .gaia/artifacts/implementation-artifacts/epic-E1-*/stories/E1-S1-*.md
Expected duration: 2 minutes per story.
Step 9 -- Pick what to work on
/gaia-sprint-plan
The sprint manager agent (Nate) selects stories from the backlog
based on priority, dependencies, and velocity. The sprint is
committed atomically to sprint-status.yaml, which
becomes your source of truth for the iteration.
Artifact produced: sprint-status.yaml
Expected duration: 2 minutes.
Step 10 -- Implement your first story
/gaia-dev-story E1-S1
The full implementation workflow: plan approval, TDD (red/green/refactor), code writing, test execution, commit, push, PR creation, CI wait, and merge. GAIA selects the right stack developer agent based on your project's tech stack and walks through each step with you.
Artifacts produced: source code, test files, PR, review reports
Expected duration: varies by story complexity.
Step 11 -- Verify Definition of Done
/gaia-check-dod E1-S1
Parses the story's DoD checklist and reports which items are checked vs unchecked. A green DoD is a precondition for the next step (running the full review battery).
Artifacts produced: none (read-only check; updates the story's DoD section indirectly via the upstream gates).
Expected duration: under 1 minute.
Step 12 -- Run all 6 reviews
/gaia-review-all E1-S1
Dispatches six review sub-agents in parallel against the just-implemented story:
code review, security review, QA tests, test automation, test review, and performance review.
Each writes a review report and updates the corresponding row of the story's Review Gate
table to PASSED, FAILED, or UNVERIFIED via
review-gate.sh. Reviews that require execution evidence (QA tests, test
automation, test review) cite the runner exit code as --execution-evidence JSON
when the Test Execution Bridge is enabled.
Artifacts produced:
code-review-{key}.md,
security-review-{key}.md,
qa-tests-{key}.md,
test-automate-review-{key}.md,
test-review-{key}.md,
performance-review-{key}.md in .gaia/artifacts/implementation-artifacts/
(per-story reviews/ subdir on the per-story layout) plus optional
evidence/*.json when the bridge is on.
Expected duration: 5-10 minutes for the parallel battery.
Step 13 -- Transition the story to done
/gaia-check-review-gate E1-S1
# (all 6 PASSED → COMPLETE)
${CLAUDE_PLUGIN_ROOT}/scripts/transition-story-status.sh E1-S1 --to review
# (verify the review gate one more time)
${CLAUDE_PLUGIN_ROOT}/scripts/transition-story-status.sh E1-S1 --to done
review-gate-check reads each Review Gate row and exits
0 for COMPLETE, 1 for BLOCKED, 2 for PENDING.
The story transitions through in-progress → review → done —
each transition is the boundary writer that atomically updates the story file frontmatter,
.gaia/state/sprint-status.yaml, the per-epic shard, and story-index.yaml.
Artifacts updated (4-surface atomic write):
story file frontmatter, .gaia/state/sprint-status.yaml,
per-epic shard at .gaia/artifacts/planning-artifacts/epics/epic-{slug}.md,
and .gaia/state/story-index.yaml.
Expected duration: under 1 minute.
Step 14 -- Repeat for the rest of the sprint
Loop Steps 10-13 for every other story in the sprint. The bridge runs unit/integration
tests across all stories as you go, and each story's reviews accumulate evidence the
sprint review will consult. Use /gaia-sprint-status at any time to see the
dashboard (sprint id, done/total counts, points burned, risk-flagged stories).
Expected duration: 1-3 days for a typical 5-story sprint.
Step 15 -- Run the end-of-sprint review
/gaia-sprint-review
Two parallel tracks: Track A dispatches Val to text-validate every story
and review against the sprint's goals using the per-goal rubric;
Track B re-runs the configured per-stack execution review.
The composite verdict
(PASSED / FAILED / UNVERIFIED) is reduced by compose-verdict.sh:
PASSED routes to /gaia-sprint-close; FAILED transitions the sprint to
correction and emits action items routed to /gaia-correct-course.
Artifacts produced:
.gaia/artifacts/implementation-artifacts/sprint-review/sprint-review-{sprint_id}-{date}.md
plus the Val dispatch sentinel at .gaia/memory/checkpoints/sprint-review-{sprint_id}-val-dispatched.json.
Expected duration: 5-15 minutes (most time is in the Val dispatch).
Step 16 -- Write the retrospective
/gaia-retro
Walks went-well, didn't-go-well, and action items. Step 7 (Val sidecar
write) is REQUIRED before finalize under
GAIA_FINALIZE_SENTINEL_REQUIRED=1 — see the
retro Step 7 note.
Action items land in .gaia/state/action-items.yaml for the next planner to consume.
Artifacts produced:
.gaia/artifacts/implementation-artifacts/retrospective/retrospective-{sprint_id}-{date}.md,
updated .gaia/state/action-items.yaml,
appended .gaia/memory/sm-sidecar/velocity-data.md.
Expected duration: 10-15 minutes (mostly facilitation).
Step 17 -- Close the sprint
/gaia-sprint-close
Atomically writes status: closed + closed_at to
.gaia/state/sprint-status.yaml, archives a copy under
.gaia/artifacts/implementation-artifacts/sprint-archive/sprint-{N}-closed-{date}.yaml,
and emits a sprint_closed lifecycle event to
.gaia/memory/lifecycle-events.jsonl. This is the sanctioned boundary-write
replacement for any manual yq -i edits on sprint-status.yaml.
Artifacts produced:
updated .gaia/state/sprint-status.yaml,
.gaia/artifacts/implementation-artifacts/sprint-archive/sprint-{N}-closed-{date}.yaml,
appended .gaia/memory/lifecycle-events.jsonl.
Expected duration: under 1 minute.
Step 18 -- Plan the next sprint
/gaia-action-items
/gaia-sprint-plan
/gaia-action-items resolves open action items from the retro (route each item
by type: clarification, implementation, process, automation). Items routed to
/gaia-create-story become new backlog stories.
/gaia-sprint-plan then selects from the (now-updated) backlog for the next
sprint, applying sizing and priority rules.
Artifacts produced:
new backlog stories under
.gaia/artifacts/implementation-artifacts/epic-{slug}/{key}-{slug}/story.md (per-story layout),
seeded next-sprint .gaia/state/sprint-status.yaml.
Expected duration: 15-30 minutes for action-items + planning combined.
You are now back at Step 10 for the next sprint. Welcome to the loop.
What to read next
- The GAIA Lifecycle -- the five-phase diagram showing how all commands connect.
- Common Workflows -- multi-command recipes for brownfield onboarding, quick fixes, deployments, and more.
/gaia-help-- context-sensitive help that suggests the best next command based on your project's current state.