BUILT BY CLOUDFLARE ON CLOUDFLARE
Tokens are cheap. Trust is not. AI made production free — and the reader stopped reading. Here's how to get them back.
For two years, teams pointed AI at their docs and walked away. The output was technically complete — covered the topic, answered the question, said nothing. The human decided what to generate, pressed enter, and left the judgment seat.
Points AI at a topic, presses enter, ships the output. Technically complete, semantically empty. The human decided what to generate but not what to say.
Learned to pattern-match the mediocre. Instinctively, the way you stop reading a publication after it burns you three times. Tokens are cheap. Trust is not.
The reader's distrust changed the economics. Generating docs got cheaper every quarter. Getting someone to read them got harder. The bar didn't fall — it rose. Quality became the scarce resource.
Pirsig's thesis: Quality emerges when the person doing the work is present in it. The mechanic who cares produces a different result. Not a different wrench — a different relationship to the work.
The mechanic can hand the wrench to a machine that turns it perfectly and indifferently. Caring didn't get easier. It got harder. The machine removed the excuse of difficulty — and with it, the forcing function.
So how do you keep a human in the seat when the machine can do the work without one?
You don't solve this with willpower. "Use AI responsibly" is the quality equivalent of "eat less and exercise more" — true, useless, and immediately forgotten under pressure. You solve it with architecture.
Drafting, formatting, rendering, publishing. Instant, tireless, indifferent. 270 tools that handle everything that isn't taste.
What to say, what to cut, what to leave out, what earns the reader's time. The only job left — and the only one that matters.
You can't publish without an editorial decision. Not policy. Architecture. The system makes the division structural.
That's the claim. Now watch it work.
19 MCP servers. 270+ typed tools. Each owns a domain — files, git, color science, workflow, rendering, publishing. The output of one becomes the input to the next.
Workspace context management. Resolves hints to paths, manages worktrees, handles all file operations. The only source of truth for where code lives.
Git primitives. Commit categorization, MR lifecycle, branch management. Takes a path from scout, executes git operations.
IDE configuration. Manages MCP server registration, agent definitions, permission rules, and plugin wiring.
Confluence integration. Three-way sync, DOM editing, page screenshots, attachment uploads, comment threads.
Page spec compiler. YAML to Confluence storage-format XHTML. 20+ block types, scheme-aware coloring, structural validation.
Diagram engine. 10 structural forms rendered as SVG. Measures labels, allocates space, degrades gracefully.
Color system. 21 OKLCH hue families × 11 shades, gamut-mapped to sRGB. Schemes, semantic roles, distribution functions.
Cloudflare Pages deployment. Project management, file and directory upload with MD5 deduplication.
Workflow engine. Phase-based tool scoping — each phase narrows available tools and directs attention through signals.
Session coordination. Tracks parallel AI sessions, detects file conflicts, manages advisory path claims.
Session analytics. Corrections and preferences that persist across conversations. Usage trends and cost tracking.
Campaign orchestration. Mission planning with dependency graphs, parallel dispatch, self-reporting completion.
Type system. Semantic annotations linking tool inputs and outputs across all servers. The composition backbone.
Log investigation. Process specs, pattern matching, KQL query generation for production debugging.
What changes when AI handles the mechanical parts of engineering? From the quality thesis through methodology, practice, and structural safety.
OKLCH color science, dark surfaces, ten diagram forms, six visual principles. The system that governs every page you see here.
The system applied to a real billing domain. Seam maps, state machines, and architectural specs. The domain is abstracted. The method is real.
The renderer that compiled this page. YAML in, HTML out. 19 block types, animated entrances, scheme-aware theming. Zero external dependencies.
This page was built by the system it describes. AI handled production. A human held the judgment seat. The quality — or lack of it — is the proof.