Skip to main content

Tech Lead Guide

This guide is for you if you make architecture decisions, set coding standards, review designs and code, or curate AI prompt engineering practices. Tech leads span both engineering and planning, with 23+ addressable assets across design, standards, review, and prompt management.

TIP

Install the HVE Core extension from the VS Code Marketplace to get all stable artifacts with zero configuration.

Your primary collections are hve-core (research and review workflows), coding-standards (language-specific rules), and project-planning (architecture decision records and planning). For clone-based setups, use the hve-core-installer agent with install hve-core coding-standards project-planning.

What HVE Core Does for You

  1. Creates architecture decision records (ADRs) capturing design rationale and trade-offs
  2. Generates architecture diagrams from codebase analysis
  3. Reviews code and pull requests against architectural guidelines and coding standards
  4. Activates language-specific coding standards automatically based on file type
  5. Builds, analyzes, and refactors prompt engineering artifacts (prompts, agents, instructions, skills)
  6. Manages research and planning workflows that feed into engineering implementation

Your Lifecycle Stages

NOTE

Tech leads primarily operate in these lifecycle stages:

Stage 2: Discovery: Research architecture, evaluate design options, gather evidence Stage 3: Product Definition: Define architecture decisions and design specifications Stage 6: Implementation: Guide implementation, enforce standards Stage 7: Review: Review designs, code, and architectural compliance Stage 9: Operations: Maintain standards, evolve architecture

Stage Walkthrough

  1. Stage 2: Discovery. Use the task-researcher agent to evaluate design options, research external patterns, and gather architectural evidence.
  2. Stage 3: Product Definition. Create architecture decision records with the adr-creation agent and generate diagrams with the arch-diagram-builder agent.
  3. Stage 6: Implementation. Guide engineers using coding standards (auto-activated by file type) and prompt engineering tools for AI artifact creation.
  4. Stage 7: Review. Run the pr-review agent for automated pull request feedback and the task-reviewer agent for implementation-against-plan validation.
  5. Stage 9: Operations. Use /prompt-analyze and /prompt-refactor to maintain and evolve prompt engineering artifacts as team practices mature.

Starter Prompts

Select adr-creation agent:

Create an ADR for adopting OpenTelemetry as our observability standard,
replacing the current custom tracing library. Cover decision drivers
around vendor neutrality and auto-instrumentation support, alternatives
like Datadog APM and Jaeger, and migration impact on existing services.

Select arch-diagram-builder agent:

Generate an architecture diagram for the event-driven order processing
pipeline. Show the message flow from API gateway through the event bus
to worker services, including the dead-letter queue and monitoring
integration. Use mermaid flowchart syntax.

Select pr-review agent:

Review the current pull request focusing on architecture alignment with
docs/architecture/ patterns, API contract consistency with existing
endpoints, test coverage for new code paths, and performance implications
of any new database queries.
/prompt-build Create a new instructions file for Python data pipeline
development. Cover pandas conventions, type hinting requirements,
virtual environment setup with uv, and testing patterns using pytest.
/prompt-analyze Analyze .github/instructions/coding-standards/python-script.instructions.md
for quality. Check frontmatter schema, applyTo coverage, instruction
specificity, and alignment with repository conventions.

Key Agents and Workflows

AgentPurposeDocs
adr-creationArchitecture decision record creationAgent file
arch-diagram-builderMermaid architecture diagram generationAgent file
pr-reviewPull request review automationAgent file
task-reviewerImplementation review against planTask Reviewer
prompt-builderPrompt engineering artifact creationAgent file
task-researcherDeep codebase and architecture researchTask Researcher
task-plannerStructured implementation planningTask Planner
doc-opsDocumentation operations and maintenanceAgent file
memorySession context and preference persistenceAgent file

Prompts complement the agents for cross-cutting workflows:

PromptPurposeInvoke
git-commitStage and commit changes with conventional message formatting/git-commit
pull-requestCreate a pull request with structured description/pull-request

Auto-activated instructions apply coding standards based on file type: C# (*.cs), Python (*.py), Bash (*.sh), Bicep (bicep/**), Terraform (*.tf), and GitHub Actions workflows (*.yml).

Tips

DoDon't
Create ADRs for significant design decisionsMake architectural choices without documented rationale
Use the pr-review agent to supplement manual code reviewsRely solely on automated review without human judgment
Let coding standards auto-activate based on file typeManually apply rules that already have instruction files
Use /prompt-analyze before refactoring AI artifactsRewrite prompts without understanding their current structure
Research with the task-researcher agent before architecture changesDesign without investigating existing patterns and constraints
  • Tech Lead + Engineer: Architecture decisions feed implementation. Tech leads set standards and review while engineers build. See the Engineer Guide.
  • Tech Lead + Security Architect: Security architecture integrates with overall system design. Threat models inform architecture decisions. See the Security Architect Guide.
  • Tech Lead + TPM: Architecture shapes product requirements and vice versa. Design decisions affect decomposition and sprint planning. See the TPM Guide.

Next Steps

TIP

See the full project lifecycle: AI-Assisted Project Lifecycle Explore prompt engineering practices: Prompt Engineering Contribution Guide Review coding standards: Coding Standards Collection


🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.