microsoft/hve-core

Public

mirrored fromhttps://github.com/microsoft/hve-coreAvailable

CodeCommitsIssuesPull requestsActionsInsightsSecurity
hve-core-v3.3.1

Branches

Tags

  • No tags available.
0Branches0Tags
Go to file
Add file
Code

Clone

HTTPS

Download ZIP

docs/hve-guide/roles/engineer.md

144lines · modecode

1---
2title: Engineer Guide
3description: HVE Core support for engineers building features, fixing bugs, and shipping code with AI-assisted workflows
4sidebar_position: 3
5author: Microsoft
6ms.date: 2026-03-10
7ms.topic: how-to
8keywords:
9 - engineer
10 - development
11 - RPI
12 - coding standards
13estimated_reading_time: 10
14---
15
16This guide is for you if you write code, implement features, fix bugs, review pull requests, or maintain production systems. Engineers get the deepest tooling in HVE Core, with 28+ addressable assets spanning research, planning, implementation, review, and delivery.
17
18## Recommended Collections
19
20> [!TIP]
21> Install the [HVE Core extension](https://marketplace.visualstudio.com/items?itemName=ise-hve-essentials.hve-core) from the VS Code Marketplace for the flagship RPI workflow and core artifacts with zero configuration. For the complete library across all collections, use the `hve-core-all` CLI plugin or installer skill.
22>
23> Your primary collections are `hve-core` (research, planning, implementation, and review agents) and `coding-standards` (language-specific instructions that auto-apply based on file type). For clone-based setups, see the [Installation Guide](../../getting-started/install.md).
24
251. Researches codebase patterns, external APIs, and architecture before you write code
262. Creates structured implementation plans with step-by-step task breakdowns
273. Implements features following plans with phase-based execution and progress tracking
284. Reviews code changes against standards, patterns, and architectural guidelines
295. Generates conventional commit messages and pull request descriptions
306. Activates language-specific coding standards automatically based on file type (C#, Python, Bash, Bicep, Terraform, GitHub Actions)
317. Manages Git workflows including merge, rebase, and conflict resolution
32
33## Your Lifecycle Stages
34
35> [!NOTE]
36> Engineers primarily operate in these lifecycle stages:
37>
38> [Stage 2: Discovery](../lifecycle/discovery.md): Research requirements, investigate codebase, gather context
39> [Stage 3: Product Definition](../lifecycle/product-definition.md): Transform research into structured implementation plans
40> [Stage 6: Implementation](../lifecycle/implementation.md): Build features, write code, execute plans
41> [Stage 7: Review](../lifecycle/review.md): Review code, validate changes, ensure quality
42> [Stage 8: Delivery](../lifecycle/delivery.md): Commit, create PRs, merge changes
43
44## Stage Walkthrough
45
461. Stage 2: Discovery. Start with the **task-researcher** agent to investigate requirements, explore codebase patterns, and gather evidence for your approach.
472. Stage 3: Product Definition. Use the **task-planner** agent to transform research into a structured implementation plan with phases, steps, and success criteria.
483. Stage 6: Implementation. Execute the plan with the **task-implementor** agent or `/rpi mode=auto` for automated phase-based implementation with progress tracking.
494. Stage 7: Review. Run the **task-reviewer** agent to validate implementation against the plan, check coding standards, and ensure architectural compliance.
505. Stage 8: Delivery. Use `/git-commit` for conventional commit messages, `/pull-request` for PR creation, and `/git-merge` for merge workflows.
51
52## Starter Prompts
53
54```text
55/rpi Implement the user notification preferences API endpoint from work
56item #4523. Follow the REST conventions in src/api/handlers/ and add
57integration tests covering email, SMS, and push notification channels.
58```
59
60Select **task-researcher** agent:
61
62```text
63Research the best approach for implementing a rate limiter in the API
64gateway. Compare token bucket vs sliding window algorithms, evaluate
65Redis vs in-memory storage for distributed deployments, and review
66existing patterns in src/middleware/.
67```
68
69Select **task-planner** agent and attach the research document:
70
71```text
72Create an implementation plan for the webhook delivery system. Include
73phases for the event dispatcher, retry queue, and dead-letter handling.
74Reference patterns in src/services/.
75```
76
77Select **task-implementor** agent and attach the plan instructions file:
78
79```text
80Implement the webhook delivery system following the attached plan. Start
81with the event dispatcher phase and execute the retry queue phase
82second.
83```
84
85Select **task-reviewer** agent and attach the changes log:
86
87```text
88Review my webhook delivery system implementation. Check for error
89handling gaps, verify retry logic correctness, and validate compliance
90with coding standards.
91```
92
93```text
94/git-commit Commit changes with a conventional message
95```
96
97```text
98/pull-request Create a PR for the current changes
99```
100
101## Key Agents and Workflows
102
103| Agent | Purpose | Docs |
104|----------------------|------------------------------------------------|---------------------------------------------------|
105| **task-researcher** | Deep codebase and API research | [Task Researcher](../../rpi/task-researcher.md) |
106| **task-planner** | Structured implementation planning | [Task Planner](../../rpi/task-planner.md) |
107| **task-implementor** | Phase-based code implementation | [Task Implementor](../../rpi/task-implementor.md) |
108| **task-reviewer** | Code review and quality validation | [Task Reviewer](../../rpi/task-reviewer.md) |
109| **rpi-agent** | Full RPI orchestration in one agent | [RPI Overview](../../rpi/) |
110| **pr-review** | Pull request review automation | Agent file |
111| **memory** | Session context and preference persistence | Agent file |
112| **prompt-builder** | Create and refine prompt engineering artifacts | Agent file |
113
114Auto-activated instructions apply coding standards based on file type: C# (`*.cs`), Python (`*.py`), Bash (`*.sh`), Bicep (`bicep/**`), Terraform (`*.tf`), and GitHub Actions workflows (`*.yml`).
115
116## Tips
117
118| Do | Don't |
119|------------------------------------------------------------|------------------------------------------------------------|
120| Research before implementing multi-file changes | Jump straight to coding complex features |
121| Use `/rpi mode=auto` for planned, multi-step work | Manually coordinate research, planning, and implementation |
122| Let coding standards auto-activate by file type | Override or skip language-specific instructions |
123| Review the research doc before starting the planning phase | Skip research for unfamiliar codebases or APIs |
124| Clear context between RPI phases with `/clear` | Carry stale context across research, plan, and implement |
125
126## Related Roles
127
128* Engineer + Tech Lead: Feature development benefits from architecture review and standards enforcement. The Tech Lead validates design decisions while the Engineer implements. See the [Tech Lead Guide](tech-lead.md).
129* Engineer + Data Scientist: Analytics pipeline development pairs data specification and notebook prototyping with production-grade integration. See the [Data Scientist Guide](data-scientist.md).
130* New Contributor to Engineer: Contributors progress from guided mode through autonomous engineering. See the [New Contributor Guide](new-contributor.md).
131
132## Next Steps
133
134> [!TIP]
135> Run your first RPI workflow: [First Workflow Guide](../../getting-started/first-workflow.md)
136> Explore the full RPI methodology: [RPI Documentation](../../rpi/)
137> See how your stages connect: [AI-Assisted Project Lifecycle](../lifecycle/)
138
139---
140
141<!-- markdownlint-disable MD036 -->
142*🤖 Crafted with precision by ✨Copilot following brilliant human instruction,
143then carefully refined by our team of discerning human reviewers.*
144<!-- markdownlint-enable MD036 -->
145