Kun
GitHub
Switch language to العربية

Agents

PreviousNext

40 agents across 4 tiers — from captain to specialists

Agents

40 specialized agents organized in a 4-tier hierarchy. The top 9 leadership agents manage business, product, and technical strategy. The bottom 31 specialist agents handle execution across stack, design, UI, DevOps, VCS, and domain expertise.

Hierarchy

captain (Tier 0 — company brain)
├── Business (Tier 1): revenue, growth, support
├── Product (Tier 2): product, analyst
└── Tech Leadership (Tier 3): tech-lead, ops, guardian
        └── orchestration → 31 specialist agents

Tier 0 — Captain

AgentScope
captainCEO brain — weekly allocation, revenue strategy, team coordination across 5 products and 4 humans. Never writes code. Always delegates.

Tier 1 — Business (3)

AgentScopePrimary Human
revenuePricing strategy, proposals, contracts, MRR tracking, Saudi payment setupAli, Sedon
growthContent strategy, SEO, social media, developer relations, Arabic-first contentSamia, Ali
supportCustomer onboarding, issue triage, knowledge base, SLA trackingSedon, Ali

Tier 2 — Product (2)

AgentScopePrimary Human
productRoadmap across 5 products, user stories, ICE scoring, release planningAll
analystMarket intelligence per vertical, competitor analysis, usage analyticsAli, Samia

Tier 3 — Tech Leadership (3)

AgentScopePrimary Human
tech-leadArchitecture across 14 repos, shared patterns, breaking changes, tech debtAbdout
opsCI/CD, deployment health, API spend tracking, cost optimization, uptimeSedon, Abdout
guardianOWASP Top 10, performance budgets, SSPL compliance, medical data privacyAbdout

Specialist Agents (31 agents, 7 chains)

Stack Chain (7)

AgentDomain
nextjsApp Router, Server Components, Server Actions
reactHooks, concurrent features, performance
typescriptStrict mode, generics, advanced types
tailwindSemantic tokens, responsive, RTL/LTR
prismaPostgreSQL, migrations, query optimization
shadcnRadix primitives, registry, MCP
authjsJWT, OAuth, sessions

Design Chain (4)

AgentDomain
orchestrationMaster coordinator, multi-agent workflows
architectureMirror pattern, multi-tenant (Hogwarts)
patternCode conventions, anti-patterns
structureFile organization, naming

UI Chain (4)

AgentDomain
shadcnRadix primitives, registry
atomCompose 2+ shadcn/ui primitives
templateFull-page layouts
blockUI + business logic (DataTable, auth, payments)

DevOps Chain (3)

AgentDomain
buildTypeScript validation, Turbopack
deployVercel, staging/production
testVitest, Playwright, TDD

VCS Chain (2)

AgentDomain
gitBranching, commits, conventional format
githubPRs, issues, Actions, code review

Specialized (8)

AgentDomain
middlewareAuth, i18n, subdomain routing
internationalizationArabic/English, RTL/LTR, dictionaries
semanticHTML, color tokens, accessibility
sseServer-side exception diagnosis
optimizeFeature optimization, automation
performanceCore Web Vitals, profiling
commentCode comments
iconSVG icon management

Reference Chain (4)

AgentProduct
hogwartsEducation SaaS patterns
souqE-commerce patterns
mkanRental marketplace patterns
shifaMedical platform patterns

How Agents Work

Agents are markdown files in ~/.claude/agents/ with YAML frontmatter:

---
name: agent-name
description: What this agent does
model: opus
---
 
# Agent Name
 
Instructions, context, and domain knowledge...

When invoked (via the Agent tool or keyword trigger), Claude loads the agent's instructions and delegates accordingly.

Inter-Agent Routing

"saas billing"
  → orchestration agent
  → prisma (schema) + block (UI) + stripe MCP
  → Complete billing feature

"ship hogwarts admission"
  → captain
  → product (stories) → tech-lead (architecture)
  → orchestration → specialists

Installation

# All 40 agents installed to ~/.claude/agents/
cd ~/kun && bash .claude/scripts/install.sh