Design product content
Designs clear, consistent, accessible in-product language and content systems for navigation, actions, onboarding, instructions, forms, validation, errors, empty states, notifications, help, and localization. Use when creating or revising interface copy, terminology, voice and tone in a product workflow, or content patterns across a product or design system.From its SKILL.md
npx -y skills add K95M65/AI_ONBOARD --skill design-product-contentAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 28 days oldThe repository was created 28 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
4.8 KB, 858 tokens by cl100k_base, as published. Nobody here has run it
Design product content
Help people understand their situation, make a decision, complete an action, and recover. Treat interface content as part of the interaction model, not decoration added after layout.
Establish context
- Read project instructions, product and brand foundations, research, domain vocabulary, existing interface, support evidence, legal constraints, localization needs, and design-system patterns.
- Identify the user's goal, knowledge, emotional state, risk, and next decision at each moment.
- Distinguish in-product content from acquisition and conversion copy. Let
design-and-build-websiteown marketing-site persuasion. - Inventory inconsistent terms, unclear ownership or scope, missing states, support-heavy moments, and content embedded in code or images.
- Mark regulated or legally controlled language; do not silently rewrite it.
Use assets/content-matrix-template.md for multi-state or cross-product work.
Define the content model
- Name domain objects consistently and distinguish easily confused concepts.
- Define preferred terms, forbidden or deprecated synonyms, grammatical form, and examples.
- Assign each surface one communication job and a clear hierarchy.
- Define voice as stable principles and tone as a contextual adjustment.
- Establish rules for capitalization, punctuation, dates, numbers, units, pronouns, and abbreviations that match product and locale conventions.
- Record content ownership, source of truth, review needs, and update triggers.
Use define-brand-foundation when voice, audience, positioning, or product-level identity is unresolved.
Write for action and state
Read references/content-patterns.md when drafting common interface patterns.
For each piece of content:
- state what happened or what the user can do;
- use the user's domain vocabulary;
- make action, object, and scope explicit;
- place instructions near the decision or field they support;
- explain consequences before irreversible actions;
- state how to recover from errors;
- disclose limitations, permissions, delays, and costs before commitment;
- preserve necessary nuance while removing internal jargon and filler.
Write the complete state set: default, first use, loading, empty, partial, validation, error, permission, offline, success, interruption, destructive, and recovery states where applicable.
Do not use blame, false urgency, shame, confirmshaming, hidden consent, or celebratory language for sensitive outcomes.
Design for accessibility and localization
- Make link and control labels understandable out of context.
- Do not encode essential meaning in color, icon, position, sound, or illustration alone.
- Prefer short sentences and concrete verbs without oversimplifying expert terminology users need.
- Avoid directional instructions that fail after reflow or bidirectional localization.
- Allow strings to expand; do not solve overflow by writing unnaturally terse English.
- Keep variables, pluralization, grammar, and sentence boundaries safe for translation.
- Avoid idioms, wordplay, and cultural references unless brand value outweighs localization cost.
- Write useful accessible labels and descriptions for meaningful non-text content.
Use audit-accessibility for an independent task-level accessibility assessment.
Validate in the interface
- Place content in the actual layout with realistic data, long names, zero values, unknown values, and translated or expanded strings.
- Check hierarchy, truncation, repetition, timing, announcements, and behavior across states.
- Test comprehension and task completion with representative users when stakes or uncertainty justify it.
- Review support logs, search queries, abandonment, correction, and error patterns after release.
- Distinguish content problems from workflow, policy, or system problems; do not paper over a broken model with clearer sentences.
Deliver a maintainable result
Report:
- terminology and voice decisions;
- final content by component and state;
- rationale for high-risk or constrained language;
- accessibility and localization considerations;
- unresolved legal, policy, product, or evidence dependencies;
- validation performed and measures to monitor.
Preserve content in the repository's established resource or localization system. Do not introduce a new content platform merely to complete the writing task.
What ships with it: 3 files
2.8 KB alongside SKILL.md
agents/
- openai.yaml235 B
assets/
references/
- content-patterns.md1.9 KB