| title | Agent Systems Catalog | |||
|---|---|---|---|---|
| description | Overview of all hve-core agent systems with workflow documentation and quick links | |||
| sidebar_position | 1 | |||
| author | Microsoft | |||
| ms.date | 2026-07-15 | |||
| ms.topic | overview | |||
| keywords |
|
|||
| estimated_reading_time | 5 |
hve-core organizes specialized agents into functional groups. Each group combines agents, prompts, and instruction files into cohesive workflows for specific engineering tasks.
| Group | Agents | Complexity | Documentation |
|---|---|---|---|
| RPI Orchestration | 1 | High | RPI Documentation |
| Code Review | 3 | Medium | Code Review |
| GitHub Backlog Management | 1 active | Very High | Backlog Manager |
| ADO Backlog Management | 2 active | Very High | Backlog Manager |
| Jira Backlog Management | 2 active | Very High | Backlog Manager |
| Project Planning | 9 | Medium-High | Project Planning |
| Security Planning | 3 active | Very High | Security Planner, SSSC Planner |
| RAI Planning | 1 active | Very High | RAI Planner |
| Data Science | 5 | Medium | Data Science |
| Experimental | 2 | Medium | Experiment Designer |
| DevOps Quality | 1 | High | Planned |
| Meta/Engineering | 1 | High | hve-builder, Documentation |
| Infrastructure | 1 | Very High | Planned |
| Design Thinking | 2 | High | Active |
The Research, Plan, Implement, Review methodology separates complex tasks into specialized phases. RPI Agent coordinates rpi-research, rpi-plan, rpi-implement, and rpi-review through durable artifacts. The self-contained rpi-challenger skill complements the lifecycle by challenging a confirmed subject through adaptive skeptical questions. See the RPI Documentation for the full guide.
A single human-gated Code Review agent provides pre-PR review on local branches. It confirms scope with you, then dispatches the perspectives you choose, functional, standards, accessibility, security, and PR, each to a thin skill-backed subagent, and merges them into one deduplicated report. A depth tier (basic, standard, or comprehensive) controls how deeply each perspective verifies the change. See the Code Review Documentation for usage guides and skill authoring.
Automates issue discovery, triage, sprint planning, and execution across GitHub repositories. The backlog manager agent orchestrates five distinct workflows with three-tier autonomy control. See the Backlog Manager Documentation for workflow guides.
Automates work item discovery, triage, sprint planning, execution, PR creation, build monitoring, and task planning across Azure DevOps projects. The ADO Backlog Manager agent orchestrates nine distinct workflows with three-tier autonomy control. The PRD-to-WIT agent translates product requirements into structured work items. See the Backlog Manager Documentation for workflow guides.
Automates issue discovery, triage, execution, and PRD-to-issue translation across Jira projects. The Jira Backlog Manager agent orchestrates workflows with three-tier autonomy control, mirroring the GitHub and ADO backlog management patterns.
Nine specialized agents for project planning activities. Includes builders for Business Requirements Documents, Product Requirements Documents, Architecture Decision Records, agile coaching, meeting analysis, network ISA-95 planning, product manager advising, system architecture review, and UX/UI design. Architecture diagrams are now delivered through the portable architecture-diagrams skill rather than a dedicated agent. See the Project Planning Agents.
Five agents handle evaluation dataset creation, data specification generation, Jupyter notebook generation, and Streamlit dashboard generation and testing.
Exploratory agents for emerging workflows. Includes the Experiment Designer for minimum viable experiment (MVE) coaching and the PowerPoint Builder for YAML-driven slide deck generation.
Agents focused on deployment reliability and build pipeline analysis.
The hve-builder skill creates, improves, refactors, reviews, and validates prompt-engineering artifacts through independent quality gates. The Documentation agent coordinates documentation audit, drift, authoring, and validation across the repository through its four modes.
Manages cloud infrastructure provisioning and configuration. Handles Bicep and Terraform deployments with validation and drift detection.
Guides teams through a six-phase security assessment covering system scoping, operational bucketing, standards mapping, security model analysis, impact assessment, and backlog handoff. The security planner agent conducts interactive sessions with structured state tracking and produces dual-platform work items for ADO and GitHub. The security reviewer agent performs automated security analysis of code changes. See the Security Planner Documentation for phase details and entry modes.
The SSSC Planner guides teams through a structured six-phase supply chain security assessment. It inventories 27 supply chain capabilities, maps against OpenSSF Scorecard, SLSA, Sigstore, and SBOM standards, performs gap analysis with adoption categories, and generates priority-sorted backlog items. Supports four entry modes: capture, from-PRD, from-BRD, and from-security-plan. See SSSC Planning for details.
Guides teams through a six-phase responsible AI assessment planning workflow covering AI system scoping, risk classification, RAI standards mapping, security model analysis, impact assessment, and review with backlog handoff. The RAI planner agent builds on security plan outputs when available and produces dual-platform work items for identified gaps. See the RAI Planner Documentation for phase details and entry modes.
The Design Thinking agents provide AI-assisted coaching through a nine-method, three-space framework for human-centered design.
| Agent | Purpose |
|---|---|
dt-coach |
Coaches teams through all 9 DT methods with session tracking |
dt-learning-tutor |
Teaches DT curriculum with exercises and assessments |
🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.