Form State Isolation
Keep long forms responsive by narrowing what re-renders when one field changes.
What it adds
Section-scoped form state with field-level subscriptions, so typing in one input does not re-render the entire form.
What your agent is told to do
5
What your agent is told to do
5-
1
Find the forms that visibly lag — the settings pages, the multi-section record editors, the pricing or configuration builders — and measure which of them re-render every field on every keystroke before changing anything.
-
2
Give each field its own subscription to its own value and its own error, so a component re-renders when the thing it displays changes and not before.
-
3
Keep cross-field rules — totals, date ranges, one-of-these-is-required, conditional sections — subscribed to the specific fields they depend on, and evaluate them against the whole form's current values rather than a stale snapshot taken when the section mounted.
-
4
Preserve section values and dirty state when a section unmounts because it collapsed, scrolled out of a virtualized list, or moved behind a step. The form owns the values; the section only displays them.
-
5
The warning shown when a user tries to leave with unsaved work belongs to Unsaved Changes Guard. This feature owns which values count as dirty and how they survive unmounting; that feature owns intercepting navigation. Do not implement a second navigation prompt here.
Edge cases it handles
7
Edge cases it handles
7- Cross-field validation must stay accurate under isolation. A rule that only re-runs when its own field changes will keep showing an end-date error after the start date was corrected, so rules must be re-evaluated when any of their inputs move.
- A collapsed, hidden, or scrolled-away section must keep its values and its dirty flag. Losing them on unmount silently discards work the user believes is entered, and it is worse than a visible error because nothing tells them.
- Fields hidden by a conditional must not contribute stale values to validation or to the payload. Decide explicitly whether hiding clears the value or merely excludes it, and apply the same rule everywhere.
- Partial saves need to know exactly which fields the user changed, so a section-level save sends only that section and does not overwrite a colleague's concurrent edit to a field the user never touched.
- Server-side validation errors arrive keyed to fields that may be inside a collapsed section, so the form must open the section, scroll to the field, and move focus rather than showing a generic failure at the top.
- Arrays of repeated rows must keep identity across reorders and deletions; keying rows by index makes the wrong row inherit another row's error and dirty state.
- An autosave or a background refresh landing mid-edit must not replace a field the user is currently typing in.
Definition of done
9
Definition of done
9- Typing in one field re-renders that field and its dependent rules, not the whole form.
- Cross-field validation produces the same result under isolation as it did when the form re-rendered wholesale.
- Collapsing, hiding, or scrolling a section away preserves its values and its dirty state.
- Conditionally hidden fields follow one documented rule for whether their values are cleared or merely excluded.
- A partial save submits only the fields the user actually changed.
- A server validation error reveals, scrolls to, and focuses the offending field wherever it sits.
- Dirty tracking is exposed for the unsaved-changes guard rather than duplicated by it.
- The feature matches the existing design system.
- No existing functionality is broken.
Related features
Prompt Versioning
Prompt Versioning
Tie every AI output to the exact prompt version that produced it.
What it does
Immutable, numbered versions of each prompt, with the run configuration recorded and every output stamped with the version used.
How it works
- 1 Make every publish create a new immutable version rather than overwriting the previous text. Editing history in place destroys the only record of what produced last month's outputs.
- 2 Capture the whole run configuration with each version, not just the wording: which model tier and parameters were used, which tools were available, and the expected output shape. A prompt that behaves differently under different settings is not one prompt.
- 3 Stamp every generated output with the version identifier that produced it, and keep that stamp with the record so an output found later can be traced back to its exact instructions.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/prompt-versioning
Retrieval Debugger
Retrieval Debugger
Show exactly which sources, chunks, and scores produced a given AI answer.
What it does
A per-answer inspector showing the query as issued, the filters applied, the candidate chunks with their scores, and what reached the model.
How it works
- 1 Capture for each answer the query as it was issued, the filters applied, the candidates returned with their scores, and which of those actually made it into the request after the context ceiling was applied.
- 2 Show results after permission filtering, with a count of how many candidates were excluded and why. Displaying the pre-filter set turns the debugger into a way to read content the viewer cannot open.
- 3 Present each scoring stage separately — keyword, semantic, and any reranking — because a chunk that ends up first overall may have been rescued by one stage after being buried by another, and a single blended number hides that.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/retrieval-debugger
AI Cost Budgets
AI Cost Budgets
Cap what AI features are allowed to spend before the bill arrives.
What it does
Monetary spending limits on AI work, scoped by workspace, feature, and time period, enforced before a run starts.
How it works
- 1 Find every place the app calls a model and route all of them through one accounting point that records estimated and actual spend against a scope. A budget that only covers the chat feature is not a budget.
- 2 Estimate the cost of a run from the size of its input before dispatching it, and refuse anything that would exceed the remaining budget on its own.
- 3 Reserve the estimate against the budget when the run starts, then reconcile to the real usage figures when it finishes, releasing whatever was over-reserved.
Copy the prompt
No account needed
Add this feature to my app:
https://addthisfeature.com/x/ai-cost-budgets
How it works
-
1
Copy the link
Grab the Markdown instruction URL for this feature.
-
2
Give it to your AI
Paste it into Claude Code, Cursor, v0, Lovable — whatever you build with.
-
3
It inspects, then implements
Your agent reads your existing app first, then adds the feature to fit it.
Works with your stack
These instructions are written to adapt. They tell the agent to detect your framework, match your existing design system, and reuse what you already have — rather than assuming a particular stack.
Need it tighter than that? Customize the feature and tell it exactly what you're running.