← ClaudeAtlas

architecture-testinglisted

Write or review tests for architectural promises: dependency boundaries, transaction atomicity, persistence mappings, stale-write detection, query budgets and API/event compatibility. Use when green tests missed a lost update or N+1, a boundary exists only in documentation, or an integration test may hide the behavior it claims to verify. Does not choose the architecture or governance thresholds (architecture-fitness-functions), replace general unit-test design, or establish production capacity (load-testing).
robsonkades/agent-skills · ★ 2 · Testing & QA · score 75
Install: claude install-skill robsonkades/agent-skills
# Architecture Testing ## Purpose and scope Turn an agreed architectural promise into a test that exercises its actual mechanism and fails for the intended violation. A green functional test does not establish transaction, dependency, concurrency or compatibility properties it never observes. This skill owns test design and implementation, not architecture selection or governance policy. Before choosing version-sensitive test APIs, inspect compiler release/toolchains, resolved framework/test-library versions, runner configuration and the JDKs used by tests and deployment. The structural example uses release 17 source compilation; its recorded run used JDK 25, not a Java 17 runtime. Match the target project rather than upgrading it to fit an example. New dependencies, preview features and toolchain upgrades require their own justification and scope. ## Workflow 1. **Name the promise and failure.** Obtain the relevant implementation, existing tests, package/contract or transaction boundary, dependency versions and build command. For data tests, also obtain engine/version, migrations, isolation and cache settings. For a budget, identify the operation and what is counted. If the promise or threshold is unknown, propose a conditional test design and request the missing contract; do not invent the architecture. 2. **Choose the smallest setup that retains the mechanism.** Use the table below. Fakes and mocks can isolate behavior but do not establish database or n