By ovargas
Enforces a disciplined development pipeline from feature ideation through pull request, integrating TDD, code review, backlog management (local or external via GitHub/Linear/Jira), structured planning and implementation, architectural validation, and automated workflow orchestration.
Document a bug report with reproduction steps, severity, and expected behavior
Knowledge check — quiz the developer on technical decisions in the current work
Create clean, atomic git commits following project conventions
Extract, define, and validate API contracts (payloads, models, events) as concrete schema files
Investigate a bug — reproduce it, trace through code, find the root cause, and document findings
Analyze and document how existing code works, tracing data flow, dependencies, and patterns.
Locate and categorize all relevant files for a given feature, component, or concept in the codebase.
Locate and categorize all relevant documents for a given topic, feature, or keyword in the project's documentation directory.
Find existing code patterns to serve as references for new work.
Analyze product ideas and features from a market, user, and value perspective.
Use when creating, reviewing, or referencing architectural decision records — defines the ADR format, three-gate creation threshold, and local/hub flavors
Use when evaluating architecture in tech-review — provides formal vocabulary (Depth, Seam, Leverage, Locality, Adapter, Pass-through) with diagnostic heuristics that produce precise, falsifiable findings instead of generic descriptions
External service backlog implementation (GitHub Issues, Linear, JIRA). Used when stack.md has backlog external with a backlog_config section.
File-based backlog implementation using docs/backlog.md (bracket markers). Default when stack.md has backlog local or no backlog field.
Abstract backlog operations interface. Load this skill before any backlog read/write operation — it defines the operations that all commands reference, delegating to the active implementation (local or external).
Modifies files
Hook triggers on file write and edit operations
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A Claude Code plugin that turns Claude Code into a virtual development team. It provides a deliberate pipeline from idea to shipped code — with agents, commands, and skills that enforce TDD, contract-first development, and quality gates at every step.
In a Claude Code session, run:
/plugin marketplace add ovargas/virtual-team
/plugin install virtual-team@virtual-team-marketplace
Start a Claude Code session and run:
/virtual-team:start
This walks you through an interactive interview to define your tech stack, project structure, and conventions. It creates stack.md (the source of truth for your project) and sets up the docs/ directory structure.
Anything you haven't decided yet gets marked as TBD — the architect agent will catch it later when a feature actually needs it.
Working on multiple repositories? You can create a hub repo to coordinate across services — shared decisions, epics, and API contracts in one place. Run
/virtual-team:start --hubin a new repo to set it up. See the command reference for details.
You only need 5 commands for daily work:
| Command | What it does |
|---|---|
/virtual-team:status | Start your day — shows what's in progress, what's next |
/virtual-team:flow <description> | Build a feature end-to-end (spec → plan → code → review → PR) |
/virtual-team:flow --fix <description> | Fix a bug end-to-end (report → investigate → fix → review → PR) |
/virtual-team:commit | Create a clean, atomic commit |
/virtual-team:handoff | End a session — captures state for the next one |
That's it. Everything else is optional.
If your project has been around for years and you're adopting this plugin retroactively, the workflow is the same — but you'll spend the first few sessions documenting what already exists rather than deciding fresh. Nothing in your code changes during onboarding. The plugin only adds documentation (stack.md, docs/) alongside your existing code.
Run this sequence the first time you open the repo:
| # | Command | Why |
|---|---|---|
| 1 | /virtual-team:start | Generates stack.md from what already exists. It reads go.mod / package.json / pyproject.toml / etc. and pre-fills the interview — you confirm what's there instead of deciding from scratch. Anything genuinely inconsistent across the codebase gets marked TBD and resolved later when a feature touches it. |
| 2 | /virtual-team:doctor | Audits stack.md against available skills. Surfaces which technologies in your stack have no coding-convention skill yet — so you know where the LLM will be working without guardrails. Read-only, never blocks. (Runs once automatically at the end of /start; re-run after editing stack.md.) |
| 3 | /virtual-team:tech-review | Baseline health reading — architecture drift, technical debt, dependency health, test coverage and security gaps. Output goes to docs/reviews/ and becomes your first backlog candidates. Skip if you don't need an audit. |
| 4 | /virtual-team:docs | Captures the tribal knowledge — setup guide, deploy guide, config reference — that's lived in people's heads. Years-old projects rarely have current docs. |
| 5 | /virtual-team:flow <new feature> | Resume normal workflow. From here, every new feature uses the full pipeline; old code stays as-is until you touch it. |
A few tips specific to legacy code:
stack.md before any feature work is a death march.tdd: recommended, not strict. Legacy code often can't be test-driven without large refactors. recommended mode (in stack.md Workflow section) lets the TDD skill adapt instead of blocking.docs/decisions/ then. Don't backfill every historical decision upfront — only record ones that are hard to reverse, surprising without context, AND a real trade-off./virtual-team:status
Shows what's in progress, what's blocked, and suggests the right command to run next.
/virtual-team:flow Add password reset via email
This runs the full pipeline in one session:
/feature → /contracts → /plan → /implement → /review + /validate → /pr
Interactive gates between each step resolve decisions and TBDs without leaving the session. If the session is interrupted, just run /virtual-team:flow again — it auto-detects where you left off.
/virtual-team:flow --fix "users can't log in after password reset"
npx claudepluginhub ovargas/virtual-team --plugin virtual-teamPersona-driven AI development team: orchestrator, team agents, review agents, skills, slash commands, and advisory hooks for Claude Code
127-agent automated development system with Agent Teams, quality gates, Bug Council diagnostics, and autonomous execution
GitHub Project Management workflow for product development: PRD creation, epic/task breakdown, TDD execution, and QA planning
Opinionated 5-phase development lifecycle for Claude Code — language-agnostic, repo-local bin/* delegation
Harness-native ECC plugin for engineering teams - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques