Inputs and forms

Inputs & Forms Skill

How this skill behaves (read first)

This is a generative skill. The default failure modes: over-asking (every field that might be useful, all required), mislabeling (placeholders as labels, jargon, unmarked optionals), and validating late (errors dumped in a modal at submit). The fix is not "apply every form pattern" — helper text, tooltips, progress bars, and multi-step splitting everywhere produce a bloated form. So this skill gates:

1. Establish context — what the form collects, why each field is needed, platform, and length.
2. Apply the always-true core — minimal fields, scannable single-column layout, persistent labels, the right input type per field, inline validation.
3. Surface the context-dependent decisions (multi-step, floating labels, helper text, pickers, smart defaults) with trade-offs.

Then it hands off to ux-heuristics-audit, ux-accessibility-audit, ux-aesthetics-audit, ux-microcopy-audit, and ux-dark-patterns-audit for validation.

---

Step 0 — Establish context before designing

Ask if not known; state the assumption if proceeding without an answer:

• What does each field buy? For every field, name who uses the data and for what. A field with no answer gets cut — the single highest-leverage form decision.
• Platform — mobile demands contextual keypads, ≥38px touch areas, a submit button that stays in sight while scrolling, one-handed reachability.
• Length & stakes — a 3-field newsletter signup and a multi-page checkout need different structure (steps, progress, error summaries).
• Sensitivity — phone numbers, payment, IDs: users ask "why do you need this?" — plan the justification (helper text) and expect dark-patterns scrutiny on anything beyond task needs.

---

The always-apply core (correct for almost every case)

Ask less

• Remove every field the task doesn't require. Shorter forms complete more; the system shouldn't ask for what it can detect or already knows (autofill known data, support browser autocomplete, auto-detect country/locale).
• Mark required and optional fields — asterisk (before the label in LTR; not pale gray) for required, "(optional)" in the label for optional.
  • ❌ Ten required fields, no explanation ✅ Five fields, one marked "(optional)", helper text on the sensitive one

Layout for scanning

• One column, top to bottom. Multi-column forms cause skipped fields, wrong-field entries, and abandonment (Baymard). Natural reading flow wins.
• Labels above inputs, outside the input, always visible. Never placeholder-as-label: it disappears on focus, strains memory, and looks like filled data. Placeholders only give example values, visibly muted.
• Group related fields (personal / billing / shipping) with headings and spacing — proximity does the structure: label-to-its-input gap < input-to-next-field gap; spacing between groups > spacing within.
• Visible input boundaries — borderless "minimal" inputs make users hunt for where to type. Input height ~32–40px, text ≥16px, labels ~13–14px, traditional restrained styling.

Pick the input type from the data

• Single-line for short answers; multi-line/text area for prose; password with a show/hide toggle; search with auto-suggest; select/dropdown only for long lists — up to ~5 options use radio buttons (single choice) or checkboxes (multiple); input masks wherever format matters (phone, date, card) — accept any way users type and format it for them; contextual keypads on mobile (numeric pad for numbers); OTP inputs that support paste and auto-fill.
• Date fields: match the picker to the task — text input + mask for known dates (birth date), calendar picker for near-future scheduling (booking).

Validate inline, kindly

• Validate on blur, not on every keystroke and not at submit. Let users finish typing; check when focus leaves the field. (Exception: live requirement checklists for new passwords.)
• Error messages sit at the erroneous input (right side or directly below — never only at the top), in plain, polite, specific language: what happened + how to fix it, ideally one line, never truncated. Color + icon + text — never color alone.
• Keep the error state visible until the input is actually valid, not just until the user starts typing.
• Prevent before you correct: constraints, forgiving formats, and masks beat error messages.

Submission is part of the form

• Primary button always in sight (below the fields; fixed on long mobile forms), labeled with the action — finish the user's sentence "I want to…": ❌ "Submit" ✅ "Subscribe" / "Place order" / "Create account".
• Button hierarchy: primary visually dominant; "Cancel"/"Back" subdued; enough spacing that the wrong button can't be hit by accident. Touch target ≥1×1cm (~38px+).
• Show all three states: loading after submit, a clear success message, and a polite, informative failure message (mind security: don't reveal which credential was wrong).

---

The context-dependent decisions (surface, don't auto-apply)

Present each with its trade-off and a recommendation tied to Step 0. Applying all of these to a 4-field form is the bloat failure mode.

Decision	Apply when	Avoid / adapt when	Default recommendation
Multi-step + progress tracker	≥3 logical steps (checkout, surveys); keeps each page digestible	Short forms — splitting adds clicks; >5–6 steps overwhelms	Single page until grouping alone stops working
Floating labels	Space-tight layouts wanting label + example in one slot	Complex/rarely-used forms where a permanent label is safer	Static labels above inputs; floating labels as a space-saving variant
Helper text	Sensitive fields ("why do you need my phone?"), nuanced requirements (password rules)	Every field "just in case" — noise that buries the signal	Only where users predictably hesitate; one line max
Icons / tooltips for guidance	Visual questions (CVV location) where a picture beats words	Tooltip-only help on mobile (extra tap, unreliable hover)	Icon first; tooltip as a supplement, not the sole carrier
Auto-suggest / incremental search	Long option lists (countries); bold the matched characters	Short lists where radio/checkboxes are faster	Auto-suggest for any list users would have to scroll
Smart defaults / auto-detection	Geo-detect country, prefill known data — fewer actions	Defaults that look like a choice the user made (consent, marketing opt-ins — dark-patterns territory)	Prefill facts, never decisions
Single full-name field	Reducing friction; more inclusive across cultures	The product must reliably split first/last name	One name field unless the data model truly needs two
Error summary at top	Long forms where multiple errors may be off-screen; each item links to its field	Short forms — inline messages suffice	Inline always; summary added for long forms
CAPTCHA / verification	Abuse is real; prefer accessible alternatives (magic links, OTP, passkeys)	Distorted-text CAPTCHAs; blocking paste in password fields	Risk-based, accessible verification; never block paste
Left-side labels	Extremely long desktop forms where vertical space is scarce	Mobile, mixed label lengths, anything responsive	Labels above inputs

---

Validate the result (orchestration)

Hand-offs name each lens by its installable skill name. Invoke one only if that skill is installed; if it isn't, this skill's own core already carries these rules — proceed without it rather than blocking.

After generating or revising, hand the result to the audit lenses rather than declaring it done. These are candidate lenses — posture is set by docs/orchestration-policy.md, or route the whole thing through ux-design-review. Here, heuristics, microcopy, and dark-patterns are Tier A (auto-run); accessibility, aesthetics, and mobile-responsiveness are Tier B (offered) — under an existing design system aesthetics is suppressed and accessibility narrows to usage; mobile-responsiveness applies only on mobile. If the user invoked this skill for one specific thing, respect that scope.

• ux-heuristics-audit (Tier A) — visibility of system status (states, loading, success), error prevention, user control, recognition over recall.
• ux-accessibility-audit (Tier B — offer; narrow under a design system) — keyboard operability and focus indicators, label/contrast minimums, aria-describedby error linking, multi-cue error states, accessible authentication.
• ux-aesthetics-audit (Tier B — offer; suppress under a design system) — spacing rhythm, grouping clarity, whether "minimal" styling destroyed input affordance.
• ux-microcopy-audit (Tier A) — label clarity, helper-text brevity, error-message tone and specificity, action-button phrasing.
• ux-dark-patterns-audit (Tier A) — over-collection of personal data, pre-checked opt-ins, sneaky defaults, required fields that should be optional.
• ux-mobile-responsiveness-audit (Tier B — offer; applies only on mobile) — when the context is mobile: keypads, touch areas, fixed submit, one-handed reach.

If the audits surface a conflict (e.g., marketing wants more fields vs. completion rate), resolve back toward the primary task: the user came to finish something — every field and interruption must earn its cost.

---

Common do/don't patterns

❌ Don't	✅ Do
Ask for everything that might be useful someday	Every field justified by who uses the data and why
Placeholder text as the only label	Persistent label above the input; placeholder = example only
Two-column field grid to "save space"	One column, logical groups, natural reading flow
Dropdown for a 3-option choice	Radio buttons (single) / checkboxes (multiple) under ~5 options
Force a strict typing format ("no dashes!")	Forgiving formats + input masks that format for the user
Validate on every keystroke, or only at submit	Validate on blur; live checklist only for password creation
Error modal at the top listing field numbers	Specific, polite message at the erroneous input (+ summary only on long forms)
Generic "Submit" on everything	Contextual action labels: "Create account," "Place order"
Submit with no feedback	Loading → success/failure states, always
Pre-checked marketing opt-in as a "smart default"	Prefill facts, never decisions (→ dark-patterns)
Ship without checking	Hand off to heuristics + accessibility + aesthetics + microcopy + dark-patterns

---

Source lessons (Uxcel)

• How to Design Forms in UIs
• Best Practices for Designing Forms
• Types of UI Inputs & When to Use Them
• Best Practices for Designing UI Inputs
• Best Practices for Styling UI Inputs
• Submitting a Form
• Showing Input Error
• Forms Accessibility