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.
Visual stability has long been treated as a performance and user experience topic. An element that moves after loading creates a poor impression, causes a misclick, or makes the site feel poorly controlled.
In the agentic era, visual stability also becomes a condition for machine action.
Why an unstable layout hurts an agent
An agent may use spatial references. It identifies a zone, associates a button with a card, locates a field, or remembers a position before interacting.
If the interface changes between observation and action, context can be lost. The target button moves. A banner is inserted. A skeleton is replaced. A component hydrates and changes size. A modal covers an area. The click becomes uncertain.
For a human, the error may be visible and corrected. For an agent, it may become a wrong action or abandonment.
CLS as an operational signal
Cumulative Layout Shift became popular as an experience metric. In an agentic path, it gains operational meaning: it indicates that the action scene is not stable.
Chrome’s Lighthouse documentation on Agentic Browsing connects stability to an agent’s ability to interact with predictable coordinates and targets. The topic is no longer only comfort. It is execution reliability.
Critical zones
Not all shifts are equal. The most sensitive ones affect:
- main navigation;
- conversion CTAs;
- submission buttons;
- form fields;
- product or service cards;
- error messages;
- modals and overlays;
- prices, variants, availability, or policies.
A decorative shift is less severe than an action shift. The audit should therefore prioritize the zones where the agent decides or acts.
Hydration and instability
JavaScript hydration is often a source of divergence. Initial HTML exposes a structure, then the client replaces, moves, or enriches it. The problem is not JavaScript itself. The problem is loss of coherence between what the agent has read and what it can act on.
A more sober architecture should ensure that critical actions remain present, named, and stable before and after hydration.
What to do
Agentic stability requires a few simple disciplines:
- reserve space for images, components, and deferred content;
- avoid injecting banners above critical actions;
- keep primary CTAs in the initial HTML;
- reduce structural changes after loading;
- test paths after hydration, not only initial rendering;
- associate each action visually and structurally with its object.
Conclusion
Visual stability becomes a condition for machine action because the agent must act on a coherent scene.
An unstable site asks the agent to recalculate the world before clicking. A stable site reduces that load, limits errors, and makes the interface more governable. CLS is therefore no longer only a UX metric. It becomes an indicator of agentic reliability.