Wizards
A Claude Code plugin marketplace for orchestrating AI agent teams to plan, build, and operate SaaS products.
Why This Exists
Building software with AI agents works best when agents have specialized roles, clear communication protocols,
and quality gates that actually block bad work. A single agent trying to be researcher, architect, engineer, and
reviewer at once produces mediocre results across the board. Specialized agents produce expert-level output in their
domain.
This marketplace packages that insight into seven Claude Code skills that spawn coordinated agent teams — each with
defined roles, structured communication, and a Skeptic who must explicitly approve all work before it advances.
What You Get
Seven slash commands across three categories:
Engineering Skills
/plan-product — Product Team
Plan what to build and why. The Product Owner coordinates a Researcher, Software Architect, DBA, and Product Skeptic to
produce implementation-ready specs.
/plan-product new user authentication
/plan-product review billing-system
/plan-product reprioritize
/plan-product # general roadmap review
Agents: Product Owner (lead), Researcher, Software Architect, DBA, Product Skeptic Output: Feature specs in
docs/specs/, updated roadmap in docs/roadmap/ Pattern: Hub-and-Spoke
/build-product — Implementation Team
Build what the Product Team specified. The Tech Lead coordinates an Implementation Architect, Backend Engineer, Frontend
Engineer, and Quality Skeptic to deliver tested, working code.
/build-product billing-system # implement a specific spec
/build-product review # review implementation progress
/build-product # resume in-progress or pick next item
Agents: Tech Lead (lead), Implementation Architect, Backend Engineer (Sonnet), Frontend Engineer (Sonnet), Quality
Skeptic Output: Working code with tests, progress notes in docs/progress/ Pattern: Hub-and-Spoke
/review-quality — Quality & Operations Team
Audit, test, and verify. The QA Lead spawns a context-appropriate subset of agents based on the task.
/review-quality security auth-module # security audit
/review-quality performance api # performance analysis
/review-quality deploy billing # deployment readiness
/review-quality regression # full regression sweep
Agents (spawned conditionally): | Mode | Agents | |------|--------| | security | Security Auditor, Ops Skeptic | |
performance | Test Engineer, Ops Skeptic | | deploy | DevOps Engineer, Security Auditor, Ops Skeptic | |
regression | Test Engineer, Ops Skeptic |
Pattern: Hub-and-Spoke
/setup-project — Project Bootstrap
One-shot setup that detects your tech stack, scaffolds the docs/ directory structure, generates CLAUDE.md, and
creates a starter roadmap. No team — runs as a single agent.
/setup-project # detect stack, scaffold everything
/setup-project --force # overwrite existing files
/setup-project --dry-run # preview without changes
Pattern: Single-Agent (no team, no skeptic)
Business Skills
/draft-investor-update — Investor Update Team
Draft investor updates from project data. Gathers metrics and milestones from the roadmap, progress files, and specs,
then drafts, reviews, and refines through dual-skeptic validation.
/draft-investor-update # current period update
/draft-investor-update Q4-2025 # specific period
/draft-investor-update status # check progress
Agents: Team Lead, Research Analyst (Sonnet), Update Drafter (Sonnet), Accuracy Skeptic, Narrative Skeptic
Output: Polished investor update in docs/investor-updates/ Pattern: Pipeline (Research → Draft → Review →
Finalize)
/plan-sales — Sales Strategy Team
Assess sales strategy for early-stage startups. Parallel analysis agents research market, product positioning, and
go-to-market, then cross-reference and challenge each other's findings before dual-skeptic validation.
/plan-sales # new assessment
/plan-sales status # check progress
Agents: Team Lead, Market Analyst, Product Strategist, GTM Specialist, Accuracy Skeptic, Strategy Skeptic
Output: Sales strategy assessment in docs/sales-plans/ Pattern: Collaborative Analysis (parallel research →
cross-referencing → synthesis → dual-skeptic)
/plan-hiring — Hiring Plan Team
Develop a hiring plan for early-stage startups. Debate agents argue for growth vs. efficiency, cross-examine each
other's cases, then dual-skeptic validation ensures fairness and strategic fit.
/plan-hiring # new hiring plan
/plan-hiring status # check progress
councilofwizards