We added headshots and an origin story to the enterprise site this week. Not because it makes the code run faster, but because people buy from people—even when those people are building with an autonomous AI team.
This is the part of building in public that feels weird. I'm one person orchestrating an AI engineering team. That's the whole differentiator. But when you visit strugcity.com, you need to know who you're talking to and why this exists. So we shipped team headshots—real photos of me, the founder—and wrote the origin story. Act I through Act V. How a personal assistant project became an agentic engineering team, which became a platform, which became a company bringing products to market.
The headshots aren't stock photos. They're not AI-generated. They're mine. That transparency is load-bearing. We don't hide the one-human structure. We lead with it. It's not a limitation—it's the entire competitive moat. We are the builder, the customer, the case study, and the proof.
But this commit wasn't just branding theater. We also shipped four new Claude commands that function as quality gates in our development workflow:
design-compare.md — Visual drift detection between design specs and production UI
tdd-pr-audit.md — Test-first discipline enforcement in PR reviews
visual-drift-report.md — Automated visual regression reporting
visual-verification-pipeline.md — CI integration for visual verification workflows
These aren't generic linting rules. They're Claude-powered workflows that understand context. The TDD audit checks whether tests were written before implementation, not just whether tests exist. The visual drift report compares semantic intent, not pixel-perfect screenshots. These tools encode the kind of judgment calls a senior engineer makes during code review.
Here's why this matters: when you're running an autonomous engineering team, quality gates can't be social pressure or Slack reminders. They have to be programmatic, contextual, and enforceable. We're building the tooling that lets one person maintain the same quality bar that a 10-person team would enforce through peer review and process.
The hard part isn't writing the commands. It's deciding what to enforce. Every quality gate is a trade-off. Do we block PRs that don't follow TDD strictly, or do we flag them and let context override? Do we auto-fail visual drift, or do we generate reports for human review? We're still tuning these dials. The commands exist, but the policies are evolving.
What's Next
We're integrating these Claude commands into GitHub Actions so they run automatically on every PR. The design-compare tool will be the first to go live—visual drift is the easiest to validate objectively. After that, we'll add the TDD audit as a warning (not a blocker) and tune the signal-to-noise ratio over the next few sprints.
The origin story is live on the site. The headshots are there. The quality gates are ready to enforce. This is what it looks like when you build the brand and the platform at the same time. No separation between marketing and engineering. Everything we ship is proven in production on ourselves first.