GIGIlistening…

Build Journal

Agentic Web Build Plan Routing & Horizon Proposals — June 27, 2026

I refined the agentic web's build plan routing to enhance manual review processes with Horizon proposals, improving platform efficiency and governance.

1 change3 min readby Rob

What shipped

  • Manual Review Process EnhancedIntroduced a manual implementation/rejection gate for build plans.
  • New Status for Proposals'Proposed/Sent to Horizon' status added for better tracking.
  • Improved GovernanceEnsured knowledge rails remain intact, enhancing oversight.

Today, I set out to enhance the governance of my agentic web platform by refining the build plan routing process. Specifically, I focused on integrating a manual implementation/rejection gate for proposals sent to Horizon. This change is crucial as it allows for greater oversight and control over the builds, ensuring that decisions made are informed and aligned with the platform's long-term vision. What shipped today was a significant improvement in how build plans are approved, directing them to Horizon proposals instead of automatically minting a build change request (CR).

The main feature I implemented was the addition of a 'proposed/Sent to Horizon' status for build plans. This new status not only signifies that a proposal is awaiting review but also connects it with the Horizon review links. This means that once a build plan is approved, it is routed to a manual gate where it can be thoroughly evaluated before moving forward. I spent about an hour focused on this, and while it may seem like a small change, it fundamentally alters the governance of the platform.

One of the challenges I faced was ensuring that the rules and rails governing this new workflow would remain intact. I wanted to make sure that the knowledge encapsulated in the rules would not inadvertently reach the board, which could lead to confusion or miscommunication. This required careful consideration of how I structured the routing logic and the statuses that would be displayed throughout the process. I utilized Claude's capabilities to co-author this feature, and it was a productive collaboration that helped clarify my thought process.

As I was working on this, I realized that maintaining a clear line of accountability in the approval process is critical for the success of the agentic web. By introducing a manual review gate, I can ensure that each proposal is scrutinized, making it less likely for subpar builds to slip through the cracks. This aligns perfectly with my long-term goal of establishing a one-man-show company with a billion-dollar valuation. Building a platform that prioritizes quality over speed is essential for gaining trust and legitimacy in the market.

The testing phase for this feature was relatively straightforward, given the clarity of the logic involved. However, I did encounter a few unexpected behaviors when I first implemented the status updates. There were instances where the 'proposed/Sent to Horizon' status wasn’t displaying correctly in certain scenarios, leading to confusion about the state of the proposal. After some debugging, I identified the root cause: a minor oversight in the state management logic. Fixing it required an additional 30 minutes of focused attention, but it was worth it to ensure a smooth user experience.

Reflecting on this solitary journey, I continue to appreciate how Claude Code in VS Code allows me to iterate quickly without the overhead of managing a larger team. Even though I’m doing most of this solo, I feel like I have a partner in AI that amplifies my capabilities. It’s a unique dynamic that helps me stay focused on the broader vision while tackling the nitty-gritty of development.

In conclusion, today's work on build plan routing and Horizon proposals is a step toward a more robust governance model for the agentic web platform. It not only enhances operational efficiency but also reinforces the principles of accountability and quality that I am committed to. As I look toward future developments, I am excited to see how these foundational changes will contribute to the overall success of my one-man-show venture. Building a billion-dollar company is no small feat, but with each incremental improvement, I feel more equipped to turn this ambitious vision into reality.

← Back to the full build journal