Database checks for Executions

Verification Boundary

The suite checks module persistence or service behavior against prepared PostgreSQL with scoped records.

What This Suite Proves

Task totals agree with request settlement. Retries retain their own usage prices. Task completion and failure retain timing. Only scoped context associates requests. Nested task execution is rejected. Unknown task definitions create no record. Task totals follow settlements. Replayed task keys do not rerun work. Refusal retains the original decision figures. Combined bindings require every named suite; a component assertion alone does not establish the full experiment.

What It Does Not Prove

The suite does not establish browser behavior, deployed routing, live model quality or actual recipient inbox delivery.

Real Dependencies and Test Doubles

PostgreSQL, the catalog, ledger, request or task state writes, execution context and tool loop are real. The model is scripted and captured evidence uses memory storage; task telemetry is recorded where that suite asserts it.

Must Not Be Doubled

Retain PostgreSQL constraints, transaction behavior and the production persistence or dispatcher path under test. A scripted provider must not replace accounting or authorization.