MODULE 17 / 18 · PART 5 · LEAD THE AGENT & THE INTERFACE
UI/UX Layout Audit & Remediation
Run an audit that produces evidence and a sequenced remediation plan.
THE SHORT ANSWER
What this module changes.
These prompts produce the audit itself, not a redesign opinion. Run them against a screenshot and the source, and demand findings, evidence and a sequenced remediation plan.
These prompts produce the audit itself, not a redesign opinion. Run them against a screenshot and the relevant source, and demand findings, evidence, and a sequenced remediation plan — the same discipline you would apply to a backend architecture review.
Dense admin dashboards in fintech, health, logistics and SACCO systems grow by accretion. Counting duplicate controls and silent failures before redesigning is what makes the cleanup defensible.
THE MODULE PROMPT LIBRARY
12 prompts for layout audit.
Copy any prompt, replace the [bracketed] slots with your own project context, and attach the files the agent needs. Use “Copy all prompts” when you want the full module in one place.
12 copy-ready prompts. Bracket text is a slot for your own context — replace it before you send the prompt.
Conduct an architectural audit of the UI/UX layout from this screenshot and the attached codebase. Prepare a comprehensive plan to eliminate toolbar fragmentation, introduce a first-class Component Palette, ensure seamless in-canvas insertion, stabilize the Inspector panel, and enforce zero-silent-catch error handling.
Why it worksNames the five workstreams up front, so the audit produces a plan rather than a list of complaints.
Audit the toolbar: list every control, where it lives, and which other surface duplicates it. Propose the consolidated structure and the migration order.
Why it worksToolbar fragmentation is measurable as duplication; counting it makes the consolidation defensible.
Design the Component Palette as a first-class surface: what it contains, how it is searched, how it is grouped, and how it stays in sync with the component registry.
Why it worksA palette that drifts from the registry becomes a second source of truth.
Audit the in-canvas insertion path: from palette click to rendered node. Where does the user lose feedback, and where can the insertion silently fail?
Why it worksInsertion bugs are almost always missing feedback, not missing logic.
Audit the Inspector panel for stability: what causes it to remount, lose selection, reset scroll, or show stale values when the selection changes?
Why it worksInspector instability is a state-ownership problem; naming the cause is the fix.
Enforce zero-silent-catch error handling across [module]: find every empty catch, every swallowed promise rejection, and every default that hides a failure. Show the replacement for each.
Why it worksSilent catches are the reason UI failures reach users as blank screens.
For each finding in this audit, give me the evidence, the user-visible symptom, the root cause, and the smallest fix — ranked by user impact over implementation effort.
Why it worksConverts an audit into a backlog ordered by value rather than by discovery order.
Separate the findings in this audit into architectural defects, implementation bugs, and design preferences. Be explicit about which is which.
Why it worksPrevents taste from being smuggled in as architecture.
For each proposed change in this plan, tell me what it breaks, what it costs, and whether it can ship independently.
Why it worksA remediation plan without a blast radius is a wish list.
Sequence this remediation plan into phases where each phase ships independently and leaves the product in a working state.
Why it worksUI refactors that cannot ship incrementally never finish.
Tell me which of these findings I should fix now, which I should schedule, and which I should explicitly accept as debt — with the reason for each.
Why it worksAn audit that does not triage becomes a document nobody acts on.
Review this audit as a hostile reviewer: which findings are unsupported by the evidence, and which recommendations are over-engineered for the actual problem?
Why it worksSelf-critique before the audit goes to a team that will challenge it anyway.
PRACTICE · 20–30 MINUTES
Try it on a project you actually own.
Audit one screen against prompt 2. Produce a table of findings with evidence, user impact, effort and ship order.
- Pick one prompt from this module that matches a real task.
- Replace every bracketed slot with context from the feature you are working on now.
- Attach only the files the agent needs. Keep secrets, credentials and personal data out of the chat.
- Ask for the artifact — a spec, diff, test, runbook or decision note — and the evidence that it works.
- Record what changed and what remains unverified in the repository.
Want the same prompts across every module? Search the full prompt library or return to the course overview.