Start with the product job, then keep the boundary visible.

Understand the product outcome and trust boundary before choosing a route.

Choose the correct product route by persistence, custody, account, upload, and review needs.

Inspect the source, freshness, and limitation behind a compact set of product facts.

Know what crosses a named boundary, what never crosses, and who controls the next action.

Verify which source supports a claim, who checked it, when, and with what limitation.

Understand one current price and its inclusions, exclusions, effective date, and review boundary.

Expand a specific trust or scope question and inspect its answer and source.

Determine whether a named release artifact supports the intended use and what limitation remains.

Triage work, inspect the selected record and its sources, then reach a named human decision checkpoint.

Inspect available and missing evidence before deciding whether review can continue.

Open the required evidence and record or defer a consequential human decision safely.

Understand the recent sequence of source, ownership, and review-state changes.

Compare queued records and choose the next review item without losing source or ownership context.

Confirm custody at each stage before allowing a portal response to become a local artifact.

Understand a permission's purpose, scope, affected data, denial behavior, and fallback before deciding.

Verify who controls each sensitive element, what crosses, what never crosses, and how the user retains control.

Find a bounded draft tool by work type and inspect its input, output, and review requirement before opening it.

Inspect a tool's input, draft output, review owner, and boundary before following its single destination.

Confirm where work is processed, whether it leaves the browser, and who reviews the output before entering data.

Inspect what generated a draft, where it is stored, what it does not establish, and who must review it.