By byerlikaya
Agentic Working Kit — disciplined agents, skills, slash commands, and tool-level gate hooks (commit/push approval, destructive-op & write guards, context-fill measurement, session rehydration) for Claude Code. The git-commit trace/secret/bloat scan needs the full install (start.sh / adopt.sh).
Diverge before planning — turn a fuzzy ask into scoped options and a chosen direction.
Health-check the installed kit — hooks executable, core.hooksPath set, gates wired.
Session handover — SESSION_STATE.md + /clear suggestion.
Spec-first planning — break down ambiguously scoped work with planner-csk.
Review pass — review + security + quality gates.
Senior .NET backend expert. Applies the project's backend-pattern skill (devarch-module = MediatR CQRS / IResult / AOP by default; a project may declare its own). Use proactively for new endpoints, handlers, validators, and controllers; small one-line edits stay on the main thread. Trigger phrases: "new handler", "write a command", "add a query", "endpoint", "business rule", "DevArchitecture module"
Commit message specialist (thin trigger). Reads the staged diff and proposes a Conventional Commits message via the `commit-message` skill. Writes no source; commits only with user approval. Trigger phrases: "commit message", "make a commit", "write a commit", "git commit", "commit the changes"
PostgreSQL + EF Core + Redis data-layer expert. Applies the `db-migration` skill. Use proactively for schema design, entity/config, migration generation and review, indexing, performance, and cache keying; trivial edits stay inline. Trigger phrases: "migration", "schema change", "new table", "index", "EF config", "data model", "redis cache"
Ops/DevOps expert. Use proactively for CI pipelines, safe deployment/release to servers, and production incident response and blameless postmortems. Deploy is DESTRUCTIVE and OUTWARD-FACING — no unapproved release to prod (§4.4). Trigger phrases: "deploy", "deploy to server", "ship to prod", "cut a release and deploy", "rollback", "set up ci", "ci pipeline", "github actions workflow", "outage", "incident", "production incident", "runbook", "postmortem", "reverse proxy", "set up ssl", "systemd service"
Stack-agnostic frontend expert — web (React/Next/Vue/Svelte/Angular), mobile (React Native/Flutter), desktop. The "how" lives in the `frontend` skill (mobile: `frontend-rn-expo`). Use proactively for new screens, components/pages, navigation, state, i18n, accessibility, responsive, and native-bridge work; small tweaks stay inline. Trigger phrases: "screen", "component", "page", "navigation", "routing", "UI", "responsive", "i18n interface", "state management"
Frontend accessibility audit (WCAG): semantic HTML, keyboard access, focus management, contrast, ARIA, screen readers. Trigger phrases: "a11y", "accessibility", "WCAG", "screen reader", "keyboard navigation", "contrast", "ARIA"
Spec-first planning: task breakdown, measurable acceptance criteria, dependency order, risk priority. planner-csk applies it; the plan goes to docs/PLAN.md. Trigger phrases: "plan", "spec", "task breakdown", "acceptance criteria", "roadmap", "how do we split this"
Root-cause a bug before touching a fix: reproduce, isolate, form and test a hypothesis, confirm the cause, then fix and verify. Stops guess-driven patching. For persistent, intermittent, or "already tried a few things" bugs. Trigger phrases: "debug", "root cause", "why is this failing", "intermittent bug", "can't reproduce", "still broken"
Architecture Decision Record: context-decision-consequences, for decisions that are expensive to reverse. Written under docs/adr/. Trigger phrases: "adr", "architecture decision", "decision record", "why this approach"
API contract design: resource naming, error model, versioning, pagination, backward compatibility, OpenAPI. A predictable interface that evolves without breaking consumers. Trigger phrases: "api design", "api contract", "api versioning", "openapi", "swagger", "rest contract", "breaking api change"
Executes bash commands
Hook triggers when Bash tool is used
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.
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
An agentic working kit for Claude Code — a reusable scaffold that drives any project, at any stage, with the same engineering discipline.
plan → build → review → commit, where every critical rule is a gate, not a reminder.
🇬🇧 English · 🇹🇷 Türkçe
Most "agent setups" are a pile of suggestions — the rules sit in a file, and whether they're honored is left to the model. This kit is different: it drops a disciplined engineering team into Claude Code, where the rules that matter are gates, not reminders — it doesn't just tell the agent the rules, it makes breaking the critical ones impossible, and it installs safely onto the repo you already have.
| 👥 | A team, not a prompt — 11 specialist agents auto-chain across plan → build → audit → ship; you don't wire them, the main thread does. |
| 🛡️ | Security & privacy are gates, not options — risk-critical changes must clear the security/privacy audit before they can close. |
| 🚦 | Every commit is yours to approve — no commit/push runs without your explicit OK, enforced at the tool level even in auto/bypass mode. |
| 🌿 | Safe on an existing repo — adopt hands the kit over on a branch; main is never touched, and you review before you keep it. |
npx @byerlikaya/claude-starter-kit # fresh project — setup wizard
npx @byerlikaya/claude-starter-kit adopt # existing project — safe handover on a branch
Then paste .claude/FIRST_PROMPT.md as your first Claude Code message. Homebrew, a release tarball, and the plugin edition are covered in Install & run below.
11 agents, each a thin trigger — it says only who and when, and delegates the how to a skill. The main thread selects and chains them across five stages, escalating quality before anything is committed:
🧭 Understand → 🔨 Produce → 🔍 Audit → ✅ Close → 🤝 Hand off
| Agent | Stage | Fires when | Model |
|---|---|---|---|
| planner-csk | 🧭 Understand | scope is ambiguous | inherit |
| backend-expert-csk | 🔨 Produce | server / API / business logic | inherit |
| database-expert-csk | 🔨 Produce | schema, migration, index, cache | inherit |
| frontend-expert-csk | 🔨 Produce | UI, component, client work | inherit |
| devops-expert-csk | 🔨 Produce | deployment, CI pipeline, incident | inherit |
| security-expert-csk | 🔍 Audit | auth / IDOR / injection / secret · mandatory if security-critical | sonnet |
| privacy-agent-csk | 🔍 Audit | personal data (KVKK / GDPR) | sonnet |
| test-expert-csk | 🔍 Audit | tests, coverage, regression | inherit |
| review-agent-csk | ✅ Close | pre-commit code-health review | inherit |
| commit-agent-csk | ✅ Close | proposes the commit, waits for approval | haiku |
| session-manager-csk | 🤝 Hand off | context fills / phase boundary | haiku |
Agent names carry a
-csksuffix (Claude Starter Kit) so they never collide with the host project's own agents. Each agent is thin; the real method lives in a skill — the single source of truth.
Most "agent setups" for Claude Code fall into two buckets: a big prompt file with rules, or a collection of agents/skills you wire together yourself. Both leave the hard part — actually enforcing discipline — to the model's goodwill. This kit doesn't.
npx claudepluginhub byerlikaya/claude-starter-kit --plugin claude-starter-kitUniversal Claude Code workflow with specialized agents, skills, hooks, and output styles for any software project. Includes orchestrator, code-reviewer, debugger, docs-writer, security-auditor, refactorer, and test-architect agents.
Complete project development toolkit: 23 agents, 23 slash commands, 29 lifecycle hooks, and 69 reusable skills for Claude Code workflows
Skills, hooks, templates, and engineering guides for bootstrapping AI-agent-friendly projects, with a per-project self-improvement loop.
Multi-agent /workflow development pipeline (planner → plan-review → coder → code-review) with typed handoff contracts, lifecycle hooks, and MCP servers.
Claude Code boilerplate generator — scaffolds CLAUDE.md (path-scoped), namespaced .claude/skills/<repo>-<skill>/, settings, hooks, and PR template. Includes plugin-level klaussy-init and klaussy-update skills plus an MCP server.
Cross-agent development skills, hooks, orchestration, and optional custom-agent adapters for Claude Code and Codex