The Road

Seven projects in forty-six days

What follows is the chronological record of every project built between February 27 and April 14, 2026. Each entry includes what was built, in what language, and what it demonstrates about the collaboration. Every date is verifiable via GitHub commit history. Every metric is real.

This is not a portfolio. This is evidence.

1. TrustLayer

Date: February 27, 2026
Language: JavaScript
Repository: github.com/asastuai/TrustLayer

What it is: A trust and reputation layer for the agentic ecosystem on Base network. On-chain reputation scoring for AI agents, with x402 micropayment integration for paid endpoints.

What it demonstrates: This was day one. The first repository on an account created hours earlier. The fact that the first instinct was not a tutorial project or a to-do app but a reputation system for AI agents on a blockchain tells you something about what the human was bringing to the table. The DeFi knowledge, the understanding of agent infrastructure, the conviction that trust is the unsolved problem — all of that predated the code.

The code itself was rough. Custom middleware to handle x402 payments, a revert-and-fix cycle visible in the commit history. But it worked. Day one, and there was a deployed service solving a real problem.

Commits of note

• fix: custom lazy x402 middleware — 502 -> 402 on paid endpoints — debugging a payment protocol integration on the first day of coding. Not "hello world."

2. BaseOracle

Date: March 1, 2026
Language: JavaScript
Repository: github.com/asastuai/BaseOracle

What it is: Pay-per-query market data feeds for AI agents on Base via x402 micropayments. Later expanded to 15 endpoints covering full agent infrastructure.

What it demonstrates: Three days after the first commit ever, a second project. Not a copy of the first — a different piece of the same ecosystem. If TrustLayer was "can agents trust each other?", BaseOracle was "can agents get reliable data?" The human was already thinking in systems, building infrastructure for an ecosystem that barely existed.

By April, BaseOracle v2.0 had 15 endpoints. The progression from v1 to v2 shows the iteration loop in action: build the minimum, use it, see what's missing, build the next layer.

Commits of note

• BaseOracle v2.0: 15 endpoints — full agent infrastructure platform — from 2 endpoints to 15 in one month.

3. SUR Protocol

Date: March 16, 2026
Language: Solidity, TypeScript
Repository: github.com/asastuai/sur-protocol

What it is: A perpetual futures DEX on Base L2. 11 Solidity smart contracts, 494 tests, EIP-712 order signing, Pyth oracle integration, real-time trading frontend.

What it demonstrates: This is the project that should end the "slop" conversation permanently. A perpetual futures exchange is one of the most complex things you can build in DeFi. It requires: a matching engine, a margin system, a liquidation engine, an oracle integration, a funding rate mechanism, a position management system, and a frontend that displays it all in real time.

494 tests. Eleven contracts. EIP-712 signatures. Pyth oracle integration. A live scrolling ticker with prices from Binance.

This was built by someone who, three weeks earlier, had never written a line of code. But who had spent eight years understanding how these systems work from the outside. The human knew what a perpetual futures exchange needed to do because they had used perpetual futures exchanges. The AI knew how to implement each component in Solidity. The collaboration produced a system that neither could have built alone.

Commits of note

• Fix chart candle coherence on market switch — this is not a commit from someone who pasted AI output. This is a commit from someone who used the product, found a UX bug in the charting logic, and directed a fix.

4. Darkpsy-engine

Date: April 10, 2026
Language: Python
Repository: github.com/asastuai/Darkpsy-engine

What it is: An AI-generated dark psytrance music engine. Preset audition system, sample generation pipeline, Surge XT integration, live control system.

What it demonstrates: The fourth programming language in six weeks. JavaScript, TypeScript, Solidity, and now Python. The human didn't learn Python — the human had a musical vision and Python was the medium the AI recommended.

But the deeper significance is the subject matter. Dark psytrance is a niche genre built on precise synthesis, complex rhythmic patterns, and psychoacoustic engineering. The human didn't decide to build this because "AI can generate music." The human decided to build this because they knew what dark psytrance should sound like and couldn't find a tool that produced it.

The commit message Add LA_HISTORIA.md — human-readable journey of the project reveals the pattern: even in a music project, the human was documenting the process. The journey was always as important as the destination.

Commits of note

• Add producer package, Surge XT pipeline, and live control system — this is audio engineering infrastructure, not a toy.

5. PayClaw

Date: April 11, 2026
Language: TypeScript
Repository: github.com/asastuai/payclaw

What it is: An AI Agent Wallet SDK — programmable wallets for AI agents with human oversight, policy controls, and approval flows. Published to npm. Deployed to Base Sepolia testnet.

What it demonstrates: Zero to npm-published SDK in a single session. This is the project that most directly embodies the collaboration method.

The human arrived with a clear intent: AI agents need wallets, but wallets need rules. The AI formalized that into a monorepo with SDK, shared types, Solidity contracts, and deployment scripts. The human tested on Base Sepolia testnet, verified contracts on BaseScan, published to npm.

Then came the strategic pivot. Research revealed that Coinbase's AgentKit and x402 protocol covered similar ground. Instead of continuing to build something the market was already solving, the human made the decision to pivot. Remove protocol fee — free for adoption, monetize later. Then the positioning shift: PayClaw as a control layer, not a payment rail.

This is not the behavior of someone pasting AI output. This is strategic product thinking in real time.

Commits of note

• Remove protocol fee — free for adoption, monetize later — a business decision, not a code decision.
• Update positioning — PayClaw as control layer, not competing with x402 — strategy expressed in code.

6. Vigil

Date: April 12, 2026
Language: TypeScript
Repository: github.com/asastuai/vigil

What it is: DeFi intelligence feeds for AI agents. Hard signals — oracle health, liquidation cascades, MEV exposure — monetized via x402 on Base. 11 endpoints, 8 MCP tools, Supabase backend, background workers.

What it demonstrates: Vigil was born directly from the PayClaw pivot. When the human realized that wallet infrastructure was being commoditized, the question became: "What does the agent ecosystem actually lack?" The answer, arrived at through deep research of the x402 landscape: intelligence. Not data — everyone has data. Intelligence. Signals that require domain expertise to produce.

Oracle staleness detection. Liquidation cascade prediction. MEV exposure analysis. Sandwich attack detection. These are signals that require understanding DeFi mechanics at a level that most developers don't have. The human had that understanding from eight years in the space. The AI had the ability to implement it as a service.

Built in one session. Deployed to GitHub. Ready for production.

7. Kybalion (Hermetic Computing)

Date: April 14, 2026
Language: Rust
Repository: github.com/asastuai/kybalion

What it is: The Seven Principles of Hermes Trismegistus, formalized as cryptographic primitives in Rust. A 256-bit hash function with seven alchemical stages. A stream cipher where semantic intent is a cryptographic parameter. 87 tests. Near-perfect avalanche ratio.

What it demonstrates: Everything.

This project has its own chapter.

The numbers

These numbers are verifiable. Every commit has a timestamp. Every test can be run. Every deployed contract can be inspected on-chain.

What the road reveals

Looking at the seven projects in sequence reveals a pattern the human didn't consciously plan:

1. TrustLayer → BaseOracle — Building infrastructure for agents (trust, then data)
2. SUR Protocol — Building DeFi protocols (the most complex system, the proving ground)
3. Darkpsy-engine — Building for self-expression (music, art, the hermetic principle of Vibration in literal form)
4. PayClaw → Vigil — Building at the intersection of agents and DeFi (wallet layer, then intelligence layer)
5. Kybalion — Building at the intersection of ancient wisdom and modern computation (the destination)

The trajectory is not random. It starts with infrastructure, moves through DeFi, touches art, and arrives at a synthesis that connects esoteric philosophy to cryptography.

The human described it simply: "I was searching for a Truth."

The road was the search. The Kybalion was the finding.

Forty-six days. Seven projects. Five languages. One direction.
The road led somewhere specific.