🚀 AI Projects & Workflows

Focus: Applying everything to real-world development.

The AI Coding Loop: Verification-Driven Development with AI Agents

The AI Coding Loop: Verification-Driven Development with AI Agents

Modern software engineering with AI is less about "perfect prompting" and more about disciplined process. While AI can generate vast amounts of code instantly, the primary challenge shifts from authorship to verification. This guide outlines a repeatable workflow to ensure AI-generated code is secure, accurate, and maintainable — supported by real-world examples from active development.

The Core Problem: The "One-Shot" Trap

The "5-second high" occurs when an AI generates a complete module from a single sentence. However, this creates a technical debt of understanding. If you cannot verify the output, you do not own the code, making it a liability rather than an asset.

Rule of Thumb: Treat AI output like code from a stranger — useful, but untrusted until proven by tests.

The Mindset Shift: Engineering over Prompting

In the AI era, your value shifts from typing speed to three core competencies:

Problem Definition: Defining the goal clearly.
Decomposition: Breaking large systems into small, testable "bricks."
Verification: Proving the result is correct via runnable constraints.

You are the Architect, AI is the Typist: You are entirely responsible for the logic, security, and data flow. The AI executes; you own the outcome.

The 7-Step Verification Loop

Never assume the agent's first output is flawless. Use this loop to guide AI incrementally rather than asking for a complete application at once:

Define the Goal: State the objective in one clear sentence.
Establish Rules: List the non-negotiable technical constraints (what must be true).
Provide Examples: Define expected Input → Output mappings.
Identify Edge Cases: List "weird" or bad situations the system needs to handle.
Request a "Small Piece": Ask for a specific function or logic gate, not the whole system.
Demand Tests: Require the AI to provide runnable assertions.
Iterate: Treat failing tests as a "flashlight" to refine your next prompt and fix ambiguities in your rules.

Precise vs. Imprecise: The Power of Constraints

Your prompts must be highly precise when it comes to rules, constraints, and edge cases. Ambiguity is the enemy of secure AI-generated code.

Example: Server-Side Cart Calculator

If you simply ask an AI to "build a shopping cart," you risk getting vulnerable client-side logic where a user could manipulate prices. Instead, define a strict trust boundary where the server is the single source of truth:

Ignore Client Prices: Never accept a price sent from the browser; use a server-side catalog.
Validation: Quantity ≥ 1; Tax/Discount ≥ 0.
Order of Operations: Apply discounts before calculating tax.
Precision: Round money to 2 decimal places (or use cents/integers).

The "Golden Rule" Prompt Template

Goal: Calculate shopping cart totals.
Rules: 
- Input: productId, qty.
- Source of Truth: Use internal PRODUCTS catalog.
- Constraints: qty >= 1; non-negative tax/discount.
- Math: Discount first, then tax; round to 2 decimals.
Examples: 2 T-shirts ($20) + 1 Mug ($12.50) = $52.50 subtotal.
Edge Cases: Unknown productId, qty = 0.
Deliver: One JS file with Node.js 'assert' tests.

Short vs. Long Prompts: The Iterative Workflow

Instead of writing one massive prompt, the most effective strategy is iterative prompting. Start with a structured, medium-length prompt to define the goal and constraints, then transition to short, highly focused commands to build and refine the output incrementally.

Example: Rapid UI Iteration via Short Prompts

During the development of a real-time events dashboard, a developer used extremely short prompts to polish the UI once the foundational context was established by the AI:

The developer prompted: "make the bar chart smaller and horizontal. different colors for channels. randomly assigned."
Because the AI already understood the established architecture, this short prompt was enough for the agent to formulate a detailed implementation plan — creating a compact horizontal layout and using a deterministic hash-to-color function so that channels kept a stable pseudo-random color across page reloads.
The developer followed up with rapid micro-prompts like "increase font contrast of labels of bar chart" and "the colors of the background and the overlays do not match the dark". The AI executed these perfectly by tuning CSS overlay tokens and applying theme-aware colors.

Actively Verifying System Logic

Verification isn't just about automated tests; it's also about actively questioning the AI's architectural decisions. When the AI added data filters to a dashboard's charts, the developer didn't just accept the code. They verified the logic by asking: "how do the filters work? do they filter visible data or data on the server?"

Only after the AI confirmed that the filters were applied securely on the server side via SQL query parameters did the developer instruct the AI to solidify this architecture: "document the changes do you?", ensuring the verified logic was permanently recorded in the project's README.

Technical Implementation (Node.js)

The following implementation demonstrates the difference between "vulnerable" code and "engineered" code — applying the 7-Step Loop to a server-side cart calculator.

// cart.js - Run with: node cart.js
const assert = require("node:assert/strict");

const PRODUCTS = {
    tshirt: { name: "T-shirt", priceCents: 2000 },
    mug: { name: "Mug", priceCents: 1250 }
};

/**
 * CORRECT: Uses trusted catalog & validates inputs
 */
function cartTotal(cartItems, discountPercent = 0, taxPercent = 0) {
    if (!Array.isArray(cartItems)) throw new Error("Invalid input");
    
    let subtotalCents = 0;

    for (const item of cartItems) {
        const product = PRODUCTS[item.productId];
        if (!product) throw new Error("Unknown product: " + item.productId);
        if (item.qty < 1) throw new Error("Invalid quantity");

        // Logic: Use PRODUCT.priceCents, NOT item.price
        subtotalCents += product.priceCents * item.qty;
    }

    const discountCents = Math.round(subtotalCents * (discountPercent / 100));
    const afterDiscount = subtotalCents - discountCents;
    const taxCents = Math.round(afterDiscount * (taxPercent / 100));
    
    return {
        subtotalCents,
        discountCents,
        taxCents,
        totalCents: afterDiscount + taxCents
    };
}

// Validation Test
const cart = [{ productId: "tshirt", qty: 2 }];
const result = cartTotal(cart, 10, 8); 
assert.equal(result.subtotalCents, 4000);
console.log("Tests Passed: Verification Successful.");

Summary

AI is a Tool, Not an Architect: You are responsible for the logic; the AI is the typist.
Failing Tests are Data: If a test fails, it reveals an ambiguity in your rules.
Fundamentals Matter More: Security, data flow, and edge-case thinking are now the primary skills of the developer.

TL;DR

This guide introduces Verification-Driven Development, a structured methodology for integrating AI into software engineering. The text argues that modern coding requires a shift from manual authorship to rigorous oversight, prioritizing problem decomposition and testing over simple prompting. To avoid untrusted code, the workflow uses a seven-step iterative loop designed to build software through small, verifiable increments. Central to this approach is treating AI as a subordinate tool while the human developer maintains responsibility for logic and security. By emphasizing strict technical constraints and edge-case identification, the generated modules become both accurate and maintainable. The developer's value now lies in high-level system design and the ability to prove that code functions correctly under pressure.