Agent instructions
Synthetic long-form coverage for Agent instructions overview.
Purpose
This page uses Agent instructions overview as a realistic documentation scenario. It is synthetic test material designed to exercise long-form rendering, navigation, search, and publication without reproducing a reference project's prose.
The structure keeps assumptions, choices, implementation notes, and verification evidence close together. It intentionally combines prose, tables, code, callouts, lists, and internal links on one page.
Publication test
This callout must survive authoring validation and hosted publication as a supported Markdown directive.
Decision matrix
| Concern | Baseline | Stress case | Evidence |
|---|---|---|---|
| Structure | One focused section | Nested navigation and long pages | Public route resolves |
| Content | Paragraphs | Tables, code, callouts, and tasks | Expected elements render |
| Delivery | Saved Current | Asynchronous hosted build | Live status and HTTP 200 |
Example
import { describe, expect, it } from "vitest";
describe("agent instructions overview", () => {
it("keeps the observable contract", () => {
expect({ status: "ready", attempt: 13 }).toMatchObject({ status: "ready" });
});
});Verification checklist
The page has one renderer-owned title.
Body headings begin at level two.
Code fences declare a language.
Confirm the hosted page after asynchronous promotion.
Detailed scenario 1
Scenario 1 for Agent instructions overview separates inputs, expected behavior, and observable failure evidence. Reviewers compare the saved source with the hosted result and check navigation, search, code rendering, and responsive layout for omissions.
Boundary coverage includes near-empty values, long identifiers, repeated operations, and reordered nodes. The result is checked through both visible behavior and delivery state rather than appearance alone.
Detailed scenario 2
Scenario 2 for Agent instructions overview separates inputs, expected behavior, and observable failure evidence. Reviewers compare the saved source with the hosted result and check navigation, search, code rendering, and responsive layout for omissions.
Boundary coverage includes near-empty values, long identifiers, repeated operations, and reordered nodes. The result is checked through both visible behavior and delivery state rather than appearance alone.
Detailed scenario 3
Scenario 3 for Agent instructions overview separates inputs, expected behavior, and observable failure evidence. Reviewers compare the saved source with the hosted result and check navigation, search, code rendering, and responsive layout for omissions.
Boundary coverage includes near-empty values, long identifiers, repeated operations, and reordered nodes. The result is checked through both visible behavior and delivery state rather than appearance alone.