In this article
AI-Assisted Project Lifecycle Overview
Overview
HVE Core supports a 9-stage project lifecycle, from initial setup through ongoing operations, with AI-assisted tooling at each stage. Every stage maps to specific agents, prompts, instructions, and skills that accelerate your work and reduce friction. Use this guide to navigate the full lifecycle and find the right tools for your current project phase.
Stage Overview
| Stage | Name | Key Tools | Guide |
|---|---|---|---|
| Stage 1 | Setup | hve-core-installer (skill), git-setup | Setup |
| Stage 2 | Discovery | rpi-research, brd-builder, security-planner, sssc-planner, rai-planner | Discovery |
| Stage 3 | Product Definition | prd-builder, requirements-author skill, adr-creation, security-planner, sssc-planner, rai-planner | Product Definition |
| Stage 4 | Decomposition | functional-planner, backlog-manager | Decomposition |
| Stage 5 | Sprint Planning | backlog-manager, backlog-management | Sprint Planning |
| Stage 6 | Implementation | RPI Agent, rpi-plan, rpi-implement, hve-builder | Implementation |
| Stage 7 | Review | rpi-review, code-review, hve-builder | Review |
| Stage 8 | Delivery | git-merge, ado-get-build-info | Delivery |
| Stage 9 | Operations | documentation, hve-builder, incident-response | Operations |
Where Are You?
Use this navigator to jump directly to the stage matching your current need.
| I want to... | Start Here |
|---|---|
| Set up HVE Core for the first time | Stage 1: Setup |
| Understand requirements or research a topic | Stage 2: Discovery |
| Create a product spec or architecture decision | Stage 3: Product Definition |
| Break work into tasks or work items | Stage 4: Decomposition |
| Plan a sprint or manage my backlog | Stage 5: Sprint Planning |
| Write code, build features, or create content | Stage 6: Implementation |
| Review code or get PR feedback | Stage 7: Review |
| Merge, release, or deploy | Stage 8: Delivery |
| Monitor, maintain, or respond to incidents | Stage 9: Operations |
Lifecycle Flow
flowchart LR
S1["Setup"] --> S2["Discovery"]
S2 --> S3["Product Definition"]
S3 --> S4["Decomposition"]
S4 --> S5["Sprint Planning"]
S5 --> S6["Implementation"]
S6 --> S7["Review"]
S7 --> S8["Delivery"]
S8 --> S9["Operations"]
S7 -.->|"rework"| S6
S8 -.->|"next sprint"| S6
S9 -.->|"hotfix"| S6
TIP
Design Thinking can feed into this lifecycle at three exit points. See the DT-RPI integration guide for details.
Stage Transition Rules
- Design Thinking Exit 1 to Stage 2: Problem statement complete (Methods 1-3) feeds
rpi-researchin Discovery - Design Thinking Exit 2 to Stage 2: Validated concept (Methods 4-6) feeds
rpi-researchin Discovery - Design Thinking Exit 3 to Stage 2: Implementation spec (Methods 7-9) feeds
rpi-researchin Discovery - Stage 1 to Stage 2: Installation complete
- Stage 2 to Stage 3: BRD complete, handoff at
docs/project-planning/ - Stage 2 to Stage 4: TPMs skip PRD when BRD is sufficient
- Stage 3 to Stage 4: PRD and ADRs finalized
- Stage 4 to Stage 5: Work items created
- Stage 5 to Stage 6: Sprint planned
- Stage 6 to Stage 7: Implementation complete,
/clearcontext - Stage 7 to Stage 8: PR approved
- Stage 7 to Stage 6: Rework needed
- Stage 8 to Stage 6: Next sprint
- Stage 8 to Stage 9: Final sprint complete
- Stage 9 to Stage 6: Hotfix needed
Coverage Notes
Implementation has the broadest tooling surface because it combines RPI, language standards, AI-artifact authoring, data science, and infrastructure work. Delivery relies mainly on prompts and auto-applied instructions, while Setup remains intentionally narrow.
Each stage page follows a consistent structure covering purpose, key activities, AI-assisted workflow, and cross-references. This format lets you navigate directly to the stage relevant to your current work and find both manual checklists and current workflow entry points.
Next Steps
TIP
Find your role-specific guide at Role Guides to see which stages matter most for your work.
🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.