All skills
hueyexe avatar

/forms-inputs-checkout

@2841c07

Use when designing, critiquing, or implementing forms, inputs, validation, checkout, registration, payment, or any data-entry flow.

Use this Skill: https://skilld.dev/gh/hueyexe/frontend-agent-skills/forms-inputs-checkout

This session only. Nothing lands on disk.

referenceschecklists.md

≈2.7k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Checklists — Forms, Inputs, and Checkout

Use these checklists before delivering a critique, redesign, implementation plan, or code for forms and checkout flows.

Discovery/context checklist

  • The user goal is stated in plain language.
  • The business goal is stated separately from the user goal.
  • The form's submission outcome is known: purchase, save, apply, subscribe, create account, send request, etc.
  • The audience is known or defaulted to general consumer users.
  • Device context is known or defaulted to responsive/mobile-friendly.
  • Locale and language requirements are known for names, addresses, phone, dates, currency, and legal copy.
  • Legal, payment, security, privacy, accessibility, or compliance constraints are identified.
  • The required data fields are separated from optional, inferable, defaultable, and deferrable fields.
  • Existing design-system or frontend constraints are known.
  • Existing evidence is noted if available: analytics, error logs, support tickets, usability findings, user research.
  • If evidence is missing, the recommendation is framed as expert heuristic guidance, not proven performance improvement.

Field necessity checklist

  • Every field has a current user or business reason.
  • Optional fields have been removed, postponed, or justified.
  • Fields needed only for account creation are deferred until after checkout unless an account is required.
  • Marketing opt-ins are separate from required consent and not defaulted in a user-hostile way.
  • Promo-code, voucher, referral, and gift-card fields are visible enough for users who need them but do not dominate checkout.
  • Sensitive fields explain why they are needed and how the data will be used/protected.
  • Data that can be inferred reliably is not re-asked.
  • Defaults are likely correct, reversible, and aligned with user interests.
  • Returning-user defaults are easy to inspect and change.

Flow and structure checklist

  • The form title matches the user's task.
  • The first question is low-friction and contextually sensible.
  • Questions appear in the order users naturally think about the task.
  • Related fields are grouped into meaningful topics.
  • Groups are separated primarily with headings, whitespace, and proximity.
  • Decorative dividers, boxes, alternating backgrounds, or extra lines are minimized.
  • The flow does not ask high-friction or sensitive questions before trust/context exists.
  • Branching paths reveal only relevant fields.
  • Dynamic fields remain visually associated with their trigger.
  • Revealing fields does not create disorienting page jumps.
  • Multi-step flows show progress only when the step sequence is stable.
  • Long forms provide save/resume or preserve entered values after errors.
  • Review steps are included for payment, legal, booking, application, or irreversible submissions.

Layout and scan-line checklist

  • There is a clear path from label to input to primary action.
  • The primary action aligns with the form's main scan line.
  • Labels and fields do not force zigzagging eye movement.
  • Two-column layouts do not break keyboard order, zoom behavior, or responsive collapse.
  • Spacing between questions is consistent and generous enough to avoid missed fields.
  • Important content is not placed at screen edges where users are likely to miss it.
  • Dense forms still preserve clear grouping and focus order.
  • Visual hierarchy makes fields, errors, help, and actions distinguishable.
  • The primary action is visually stronger than secondary actions.
  • Destructive secondary actions are separated, lower emphasis, and recoverable.

Labels and microcopy checklist

  • Every field has a visible, durable label.
  • Label text is short, natural, and uses the user's language.
  • Labels remain visible after typing.
  • Placeholder text is used only as supplementary example/help, not as the only label.
  • Capitalization is consistent and not all-caps except for rare warnings.
  • Required/optional markers are text-based or explained with an accessible legend.
  • Help text explains only what is not obvious from label/control/format.
  • Help text is adjacent to the relevant field or placed in a consistent help area.
  • Long instruction blocks before the form are removed or shortened.
  • Copy does not rely on color, direction, or visual position alone.
  • Error copy states what happened and how to fix it.
  • Success copy states what happened and what the user can do next.

Control selection checklist

  • Radio buttons are used for small mutually exclusive option sets.
  • Checkboxes are used for independent selections or a single optional yes/no choice.
  • Select boxes are avoided for short option sets.
  • Long lists use accessible search/autocomplete/combobox behavior when appropriate.
  • Text input is used only when free-form entry is truly needed.
  • Field width provides an affordance when answer length is known.
  • Arbitrary varying field widths are avoided.
  • Input examples and masks do not make easy answers harder.
  • Users can click/tap both the control and its label to select an option.
  • Mobile keyboard type supports the expected answer.
  • Flexible formatting is accepted for common human-entered values such as phone, card, dates, and postal codes when possible.

Validation and error checklist

  • Validation is not shown while the user is still typing a normal in-progress answer.
  • Inline validation is used for high-error, strict-format, unique, availability, limit, or suggestion fields.
  • Character counters update live when limits matter.
  • Autocorrection/normalization happens after entry, not mid-entry.
  • Server validation preserves user-entered values.
  • Long forms provide a top-level error summary.
  • Error summary identifies the problem and links/focuses to invalid fields.
  • Field-level errors are placed near the field and explain the remedy.
  • Error fields use more than color: text, icon/shape, border, and programmatic invalid state.
  • Red/warning icons are reserved for errors or urgent warnings.
  • Focus moves predictably after submit errors.
  • Users can recover without losing data.
  • Duplicate submissions are prevented after clicking submit.
  • Loading/processing state is visible and announced where needed.

Checkout checklist

  • Guest checkout or account-after-purchase is supported unless account is truly required.
  • Checkout removes unrelated navigation, promotions, and offramps.
  • Order summary remains available with item count, item details, total, fees/taxes when known, shipping, and editable cart link.
  • Contact/email is requested before information needed for receipt/status.
  • Country/delivery information appears before delivery options that depend on location.
  • Delivery, billing, and payment fields appear in a sequence that matches the user's task.
  • Billing same as shipping is defaulted only when likely and easy to change.
  • Promo-code entry is available but secondary.
  • Payment/security reassurance is specific and truthful.
  • The final button label communicates the consequence: pay, place order, subscribe, submit application.
  • The review step shows final total and critical terms before purchase.
  • Post-purchase success confirms the order and provides receipt, tracking, account creation, or next actions.

Accessibility checklist

  • Each control has an accessible name.
  • Labels are associated using native label semantics.
  • Radio/checkbox groups use fieldset and legend or equivalent accessible grouping.
  • Help and error text are associated with controls.
  • Required/optional status is available to assistive technology.
  • Invalid fields expose an invalid state programmatically.
  • Error summary is focusable and announced.
  • Keyboard order matches visual/task order.
  • All actions, help triggers, overlays, and custom controls work without a mouse.
  • Focus states are visible.
  • Focus is managed after opening/closing overlays and after submission errors.
  • Text and controls scale without overlap or clipping.
  • Contrast is sufficient for text, borders, focus, and error states.
  • Color is not the only cue.
  • Motion and animation are nonessential, brief, and not flashing.
  • Timeouts can be extended or avoided without data loss when possible.
  • The form is testable with screen reader, keyboard, touch, and zoom.

Frontend implementation feasibility checklist

  • DOM order matches the intended reading and tab order.
  • Native controls are used unless a custom control is justified.
  • Custom controls implement keyboard interaction, roles, names, state, focus, and disabled/expanded/selected semantics.
  • Form component APIs support label, hint/help, error, optional/required text, described-by, and status.
  • autocomplete tokens are included for common personal, address, and payment fields.
  • type, inputmode, autocapitalize, and spellcheck are chosen intentionally.
  • Error IDs are stable and can be referenced by aria-describedby.
  • Client validation and server validation share clear field keys.
  • The implementation preserves values after validation failure.
  • Async validation handles loading, success, failure, and stale responses.
  • Submit handling prevents duplicates and supports retry on recoverable failure.
  • Layout uses responsive CSS without changing semantic order.
  • Browser differences in form controls are accepted or normalized through tested design-system styles.
  • The implementation is maintainable with tokens for spacing, color, typography, border, focus, and state.

Critique response checklist

  • The critique names the likely user goal and any assumptions.
  • Findings are ordered by impact, not by visual location alone.
  • Each finding explains the user problem and practical fix.
  • Accessibility defects are clearly labeled and not treated as optional polish.
  • Conversion recommendations are framed as hypotheses when no data is provided.
  • The critique avoids vague advice like "make it cleaner" without specifying the form behavior to change.
  • The critique includes at least one quick win and one higher-impact structural recommendation when available.
  • Tradeoffs and exceptions are stated.
  • The response does not over-ask; it asks only for missing context that changes the recommendation.

Final response checklist

  • The recommended default is explicit.
  • Exceptions are stated briefly.
  • Any user questions are specific, necessary, and include a recommended default.
  • The answer includes frontend/accessibility implications when implementation is in scope.
  • The answer distinguishes facts from assumptions.
  • No long copyrighted excerpts from source material are included.
  • The response gives the user something actionable: a revised flow, field list, wireframe description, component spec, critique, code guidance, or test plan.

Source: SKILL.md on GitHub

No alerts17d3 checks · Risk SAFE
  • Gen Agent Trust Hub17d

    The skill is a comprehensive tool for designing and critiquing forms and checkout flows. It is safe from direct malicious code but possesses a low risk of indirect prompt injection because it processes untrusted external data (such as code and design descriptions) to generate critiques and implementation guidance.

  • Socket17d

    No alerts

  • Snyk17d

    Risk: LOW · No issues

Signed by skilld at 2841c07. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last month.

Steadyupdated 2 months ago
metadata
{
  "author": "hueyexe"
}

README badge

README badge for hueyexe/frontend-agent-skills/forms-inputs-checkout