Created by Yasir Ikram • Yaz Inc

A Growing Arsenal of Agent Skills.Uncompromising Software Architecture.

The official Yaz Inc capability library: a centralized, evidence-based collection of reusable AI agent workflows for development, testing, Laravel, security, AI systems, databases, data operations, and analytics.

Curated and recommended by Yasir Ikram, backed by 15+ years of application development experience.

Explore all skills Start using the toolkit Visit www.yaz-inc.com
  • Human-readable
  • Agent-ready
  • Security-aware
  • Cross-project
Growing skill libraryFocused workflows ready for agents
Validated structureNamed, indexed, and traceable
Cross-project useApply without copying client data
Central sourceImprove once, reuse everywhere

Why this toolkit exists

Stop rediscovering the same solution.

Every completed project teaches repeatable lessons. The toolkit turns those lessons into focused instructions that another AI agent can discover, understand, and apply safely.

01

Choose the right workflow

Start from a proven checklist for the task instead of improvising architecture, tests, security controls, or data operations.

02

Preserve delivery quality

Each skill carries safety boundaries, evidence expectations, validation steps, and completion criteria into the next project.

03

Build organizational memory

Useful project experience is generalized, stripped of secrets and client details, and saved back into one canonical library.

Complete capability catalog

Find the skill your project needs.

Search by name, purpose, technology, or workflow. Open any skill to read its full instructions and source references.

Skill names and instruction files remain in English so invocation works consistently across agents.

Showing the complete skill library

Smart implementation playbooks

Start from the right solution pattern.

Four carefully selected playbooks turn useful technologies into clear starting points. Choose an outcome, understand the approach, then copy a production-aware starter prompt.

Curated guidance, not automatic approval. Review licensing, privacy, security, operating cost, and production fit before adoption.

Practical setup guide

Install once. Use with any supported agent.

One command installs the complete skill library. You do not need to copy folders or memorize platform paths.

One installer command with a simple target selector. Use -Target all, or choose codex, cursor, claude, or antigravity.

1 One-command setup

Clone and install the complete toolkit

Run this once in PowerShell. The installer safely places the full skill library in the supported user-level locations.

git clone https://github.com/Yaz-inc/yazinc-ai-toolkit.git; cd yazinc-ai-toolkit; powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\install-toolkit.ps1 -Target all

Already cloned? Run git pull; powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\install-toolkit.ps1 -Target all to update every managed copy.

3 Use a skill

Ask by name and outcome

This plain-language form works across agents. The agent reads the skill instructions and applies them to your active project.

Use the test-software-app skill to audit this application before launch.
Use the ui-ux-product-design skill to improve this dashboard and verify accessibility.
4 Let the agent choose

You do not need to know every skill

Describe the result you need and ask the agent to select the best matching toolkit skill before starting.

Inspect my request, select the most relevant Yazinc AI Toolkit skill, state which one you selected, then complete the work.

Direct discovery is also available through /skills in Codex and Antigravity, /skill-name in Claude Code, and the skills menu in Cursor.

5 Optional project-only install

Keep specialized skills inside one project

From the toolkit folder, install the library into a specific project instead of your user profile.

powershell -NoProfile -ExecutionPolicy Bypass -File .\scripts\install-toolkit.ps1 -Target all -Scope project -ProjectPath "C:\path\to\your-project"
01DiscoverSearch this catalog
02ReadOpen the full SKILL.md
03ApplyWork within its safety rules
04VerifyCollect completion evidence
05ImproveReturn reusable learning

Close the learning loop

Turn project experience into the next reusable skill.

When a workflow will be useful again, load $extract-to-toolkit. Generalize the lesson, remove secrets and client-specific data, record provenance, then update the catalog and registry together.

Read the contribution workflow Open the extraction skill

Non-negotiable rules

  • Never commit secrets, credentials, or client data.
  • Keep one skill focused on one repeatable job.
  • Record the source project and extraction date.
  • Update INDEX.md and projects/REGISTRY.md together.
  • Verify behavior before claiming a skill is ready.