Engineering playbooks for modern stacks & AI security.
Curated by Umar Farooq. Tested full-stack project setups with pre-wired AI agent rules (transpiled to 6 dialects) alongside comprehensive threat models and verified security masterclasses.
How It Works
From repo clone to autonomous AI generation with zero architectural drift.
Clone a Kit
Pick your preferred stack from the hub. Every complete kit includes pre-configured Docker containers, database environments, and single-source rules.
Bootstrap Your Agent
Feed the kit's agent_bootstrap_prompt into Claude Code, Cursor, Copilot, Cline, Codex, or Antigravity to anchor the agent's context.
Build with Safety Rails
Your agent generates code aligned with pre-verified skills, DESIGN tokens, and zero plaintext secret policies. Gitleaks and Semgrep catch violations on commit.
Featured AI Stack Kits
Pre-wired with single-source rules transpiled for Cursor, Claude Code, Codex, Copilot, Antigravity, and Cline.
Laravel 13 + Filament 4 + PostgreSQL
Enterprise administration hub with strict UUID v7 schemas, BOLA authorization policies, zero plaintext credentials, and single-source agent rules.
Laravel 12 + Livewire 4 + PostgreSQL
Full-stack reactive web architecture using Livewire Volt single-file components, property locking, real-time validations, and Alpine.js micro-interactions.
Laravel 12 + MongoDB (NoSQL)
Document store design with mongodb/laravel-mongodb Eloquent drivers, strict MongoDB injection protections, and polymorphic activity logging.
Next.js 15+ + Supabase + shadcn/ui
Modern TypeScript stack featuring Server Actions with Zod validations, Supabase Row-Level Security (RLS), and tailwindcss v4 token integration.
Security Masterclass Curriculum
22 in-depth lessons covering web application hardening and cutting-edge AI security defenses.
Laravel & AI Security Track
Complete security masterclass: prompt injection defense, Passkeys, RAG multi-tenant isolation, agent tool sandboxing, and Saudi PDPL compliance.
Production Security Checklist
Interactive 28-point pre-flight checklist. Track your progress across authentication, API gates, LLM pipelines, and server headers with browser persistence.
Security Masterclass Hub
Central catalog of all security engineering tracks, threat modeling playbooks, interactive compliance tooling, and multi-cloud security baselines.
Why Trust This
No phantom files, no hallucinated imports, no unverified claims.
Automated CI Security Scanners
Every kit runs Gitleaks to enforce zero-plaintext credential invariants and Semgrep AST scanning on every commit. Insecure patterns fail CI immediately.
Single-Source Zero-Drift Rules
Rules are written once in Markdown and compiled deterministically into 6 native formats: Cursor MDC, Claude Code, Copilot, Cline, Codex, and Antigravity.
Published Benchmark Evals
We validate each kit with evals/runner.py against real code scaffolds, asserting model compliance on BOLA, NoSQL injection, and RLS gates.
Got Questions?
Everything you need to know about TechTalks Stack Kits and security playbooks.
What is a stack kit? ↓
A stack kit is a pre-configured production blueprint combining a modern backend, frontend, and database with pre-wired AI agent rules (Cursor, Claude Code, Copilot, Cline, Codex, Antigravity), runnable starter containers, security invariants, and benchmark evals.
Which agents are supported? ↓
Every kit provides single-source agent rules compiled to 6 native dialects: Cursor (.cursor/rules/*.mdc), Claude Code (CLAUDE.md), GitHub Copilot (.github/copilot-instructions.md), Cline / Roo (.clinerules), OpenAI Codex (CODEX.md), and Antigravity (.agents/rules/*.md).
Is it free and open source? ↓
Yes. TechTalks Stack Kits is released under the permissive MIT License. You can freely clone, modify, and use all kits and security guides in commercial or private projects.
Can I contribute a kit? ↓
Yes! We welcome community contributions. Every new kit must include a valid kit.yaml manifest, starter container, single-source rules, skills, security scanners, and benchmark tasks evaluated by evals/runner.py.
How is this different from a starter template? ↓
Traditional templates only provide source files. A TechTalks Stack Kit includes five unified layers: a physical starter container, transpiled agent rules for 6 AI coding tools, reusable agent skills, an opinionated design system (DESIGN.md), and automated security scanners (Gitleaks + Semgrep).
Umar Farooq
Author & MaintainerSenior Full-Stack & Security Engineer • Founder of World Web Tree (worldwebtree.com)
Over 5 years of production experience architecting high-traffic Laravel ecosystems, PostgreSQL data pipelines, and AI security guardrails. Dedicated to providing open, reproducible standards for autonomous coding agents.
Build with Us
Have a favorite production stack? Help turn blueprints into complete kits or share reusable skills with the community.