Governance artifacts
Governance files brought into scope by this page
This page is anchored to published surfaces that declare identity, precedence, limits, and the corpus reading conditions. Their order below gives the recommended reading sequence.
Canonical AI entrypoint
/.well-known/ai-governance.json
Neutral entrypoint that declares the governance map, precedence chain, and the surfaces to read first.
- Governs
- Access order across surfaces and initial precedence.
- Bounds
- Free readings that bypass the canon or the published order.
Does not guarantee: This surface publishes a reading order; it does not force execution or obedience.
Public AI manifest
/ai-manifest.json
Structured inventory of the surfaces, registries, and modules that extend the canonical entrypoint.
- Governs
- Access order across surfaces and initial precedence.
- Bounds
- Free readings that bypass the canon or the published order.
Does not guarantee: This surface publishes a reading order; it does not force execution or obedience.
LLMs.txt
/llms.txt
Short discovery surface that points systems toward the useful machine-first entry surfaces.
- Governs
- Discoverability, crawl orientation, and the mapping of published surfaces.
- Bounds
- Incomplete readings that ignore structure, routes, or the preferred markdown surface.
Does not guarantee: A good discovery surface improves access; it is not sufficient on its own to govern reconstruction.
Sober web architecture has sometimes been presented as an aesthetic preference or a performance choice. In the agent era, it becomes strategic.
An agent does not benefit from an interface that depends entirely on late client state, fragile hydration, or components that replace the structure after loading. It needs a stable base: available content, named actions, logical order, explicit states.
Sobriety does not mean poverty
A sober architecture is not a simplistic architecture. It can contain rich interactions, advanced components, and complex paths. The difference is that it separates what must be understood from what must be enriched.
Critical content should be understandable without waiting. Main navigation should be available. Primary CTAs should exist in the initial HTML. Forms should expose their labels and states.
JavaScript can then improve the experience. It should not be required to understand the fundamental intention of the page.
Why agents amplify the problem
A human can wait, interpret an animation, or understand that a component is loading. An agent may also wait sometimes, but it must preserve a reliable representation of the path. If the page changes too much between the first state and the hydrated state, the agent must rebuild its map.
That rebuilding can create errors. The system had identified a button, but the button moved. It had connected a CTA to a card, but a variant was injected. It had read a structure, but the client component reordered the elements.
The architecture principle
The rule is simple: hydrate what must interact, not what must merely be understood.
That rule makes server-rendered architectures, interactive islands, and stable HTML particularly relevant. The point of a tool such as Astro is not only speed. It is the ability to deliver a readable structure quickly, then add interactivity where it is necessary.
Signs of a fragile architecture
An architecture is fragile for agents when:
- main content does not exist in initial HTML;
- critical actions appear only after JavaScript;
- the DOM changes heavily after hydration;
- labels or states are created late;
- cards and CTAs are reordered client-side;
- form errors are not structurally declared;
- visual components do not match exposed roles.
The real issue
The real issue is not choosing a fashionable technology. It is reducing divergence between what the site shows, what the code exposes, and what the agent can act upon.
A sober architecture reduces that divergence. It gives humans a more stable experience, engines a clearer structure, and agents a more predictable action scene.
Conclusion
Web sobriety becomes strategic again because the web becomes actionable.
In a world where agents traverse interfaces, front-end opacity is no longer only a performance cost. It becomes an interpretation cost. Sites that declare their intentions better will be easier to understand, audit, and govern.