apexCustom AI Agents
Autonomous assistants tailored to your workflows, tools, and data.
01 · Introducing apexCustom AI solutions
Custom AI solutions
and AI education
for modern teams.
Anybody (including us) can build AI tools. We make sure your teams actually know how to take advantage of them.
What we do
- → Custom AI agents
- → Workflow automations
- → Team skills libraries
- → AI education
02 · First ship
Our first open-source tool ai-agent-skills hit 998 stars in four months.
998
GitHub stars
11K+
npm installs
64
Curated skills
5
Shelves
03 · Services
Some of what we build.
Custom AI agents, workflow automations, prototypes, mobile apps, and the team skills libraries we’re known for. Every scope written in plain English before we start.
Workflow Automations
Reduce manual work with reliable automations across your core systems.
AI Prototypes
Rapid prototypes to validate feasibility and ROI quickly.
iOS / Android Apps
Mobile apps with AI at the core, built for iOS and Android.
Team Skills Libraries
Curate and ship a managed Claude Code skills library your whole team installs with one command.
Custom AI Agents
Workflow Automations
AI Prototypes
iOS / Android Apps
Team Skills Libraries
03 · EducationTeaching teams to
Teaching teams to
use AI properly.
A hands-on curriculum we’re writing for teams adopting Claude Code. Drop your email to know when each chapter lands.
In production
Claude Code Mastery
A practical walkthrough of the workflows we actually use day-to-day, with the tooling, shortcuts, and prompt patterns that compound.
- 01Setting up Claude Code for real work
- 02Working alongside Cursor
- 03Scaling prompts into playbooks
In production→
Building AI Agents
Build one from scratch. Wire it to real tools, test the edges, and learn where token costs actually leak.
- → Agent from zero
- → Tool use & evals
In production→
Agent SDK Introduction
Anthropic's Agent SDK from the ground up. Tools, skills, memory, sessions, and the trade-offs between each.
- → Why the Agent SDK exists
- → Skills, tools, and memory
06 · Live sessionsLive deep dives
Live deep dives
on the tools we use.
Real workflows, real projects, no fluff. Drop your email to get notified when the first one goes live.
Cursor, properly
The underused Cursor features that actually move the needle: debug mode, the built-in browser, workspaces, skills, MCP servers, and model selection strategy.
Claude Code
Anthropic's CLI tool for AI-assisted development, the way we use it. Setup, workflows, and the power-user patterns that compound.
Claude Cowork
A hands-on look at collaborative AI workflows for teams. How to integrate AI into a team's daily development process without it becoming chaos.
Get notified for the first session
One email when we go live. No drip sequence, no spam.
04 · ShippedReal tools,
Real tools,
in production.
Live products being used today. Built solo or with a small team.
ai-agent-skills · v4.2.0
$ npx ai-agent-skills init-library team-hq
✓ Team workspace ready · 5 shelves · shareable via git
Frontend10 skills
Backend5 skills
Mobile24 skills
Workflow11 skills
Agent Engineering14 skills
998
★ stars
11K+
npm installs
112
forks
AI Agent Skills
Team skillsShipped
The package manager for team skills libraries. Curate a managed workspace, add skills from any source, then share with the team via git. Launched Dec 2025, now the backbone for building team agent libraries across Claude Code.
JavaScript · Node.js · Dec 2025 · PresentView on GitHub →
Livehalaali.app · iOS · Android
Halaali
MobileShipped
The all-in-one Muslim lifestyle companion. Halal restaurants, grocers, butchers, masjids, prayer times, and Qiblah compass in one app. Entire stack built solo. Certified location data through partners in Canada and Mauritius.
400+
users
1,300+
locations
180+
mosques
React Native · Supabase · PostGIS · Mapbox · 2024 · PresentVisit halaali.app →
Syntera
FintechLaunching 2026
A new class of financing for small and medium businesses. Something remarkable being built quietly. Full launch in 2026.
Next.js · SMB finance · Launching 2026Visit syntera.xyz →
05 · The teamEngineering speed.
Engineering speed.
Product thinking. One roof.
Founder-led engagements. No account managers, no junior devs on your project. You talk directly to the people writing the code and shipping the work.
AS
Abdul Moiz Shahzad
EngineeringFull-stack engineer. Creator of ai-agent-skills (998 stars, 11K+ installs). Builds end-to-end from scoping to deployment and handoff.
AP
Ahmed Patel
Product & OperationsKeeps scope honest and delivery accountable. Brings clarity, pricing discipline, and the eye for what gets shipped versus what gets cut.
08 · ResourcesSharing what
Sharing what
we’re learning.
AI moves fast. We share what we discover every week so other teams can build responsibly.
In draft
Read the draft →Building team agent skills libraries
How to curate, share, and maintain a team skills library with ai-agent-skills. The patterns we used to build the one that hit 998 stars.
Essay
In progress →Designing honest AI agents
How to design, test, and deploy agents that automate real workflows without quietly drifting or burning tokens.
Notebook
In progress →Lessons on crafting software people actually use
Hard-earned notes on design, craft, and the small decisions that separate shipped products from abandoned ones.
09 · FAQQuestions we
Questions we
get asked a lot.
Direct answers, not marketing fluff.
Custom AI agents, workflow automations, internal tools, and the team skills libraries that make Claude Code useful across a whole team. If you have a shipping-oriented problem that AI can help with, that's our lane.
Founders and teams who want to build with AI but don't want to hire a full ML team. You have an idea or a workflow, you want it shipped fast, and you want to understand and own what you're getting.
No-code rebuilds, fractional-CTO retainers, or projects with no defined problem. We scope aggressively, so if the scope is genuinely undefined we'll tell you upfront instead of billing hours to figure it out.
Hourly with a cap, or fixed scope where the brief is clear. We prefer capped hourly for AI work because the shape of the problem changes as you learn. We'll tell you what we expect it to cost before we start.
The founders. No account managers between you and the engineer writing the code. You talk directly to Moiz and Ahmed.
Weeks, not months. First usable version in 1-2 weeks for most scopes. We ship the riskiest assumption first, validate it, then build on top. You see working code every week.
Get started
Book a 15-min scope call →Ready to ship?
Tell us what you’re building. 15 minutes, no pitch deck, just questions. We’ll tell you upfront if it’s not a fit.
- → You own everything we build
- → No long-term lock-in
- → We respond within 24 hours