Agent skill

Product Design

by mblode in mblode/agent-skills

Specifies interaction choices, action scope, reversibility, recovery, and reachable states.

MITAuto-check passedFrontend & Design

Install Product Design

skills CLI
$ npx skills add mblode/agent-skills --skill product-design -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install mblode/agent-skills product-design --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/mblode/agent-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/product-design .claude/skills/product-design && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
product-design
GitHub stars
143
Token cost
~2.8k tokens
SKILL.md length
1,322 words
Files
7 (incl. references)
Skills in repo
28
Repo updated
First seen
Licence
MIT

At a glance

Specifies interaction choices, action scope, reversibility, recovery, and reachable states.

  • Works in 6 steps: The user's explicit goal and constraints. → Verified user and product evidence, and… → Project-canonical guidance: AGENTS.md or… → …
  • Asked to design the flow
  • SKILL.md covers Routing boundary, Request modes, Workflow and Decision authority, plus 4 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Product Design is an agent skill from mblode/agent-skills. Specifies interaction choices, action scope, reversibility, recovery, and reachable states. Use when asked to "design the flow", "should delete be undoable", "choose the control", or "review this product decision".

Its SKILL.md is about 2.8k tokens, which your agent loads only when the skill is triggered. The skill folder holds 8 other files, including reference files (for example `evals/evals.json`, `references/interface-quality.md` and `references/naming-and-copy.md`).

It sits in Frontend & Design. The repository describes itself as: Nobody ships AI slop on purpose. These skills make sure you don’t. The licence is MIT.

When your agent uses it

  • Asked to design the flow
  • Should delete be undoable
  • Choose the control
  • Review this product decision

Example prompts

  • “design the flow”
  • “should delete be undoable”
  • “choose the control”
  • “/product-design”

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

  1. The user's explicit goal and constraints.
  2. Verified user and product evidence, and what the system actually does.
  3. Project-canonical guidance: AGENTS.md or CLAUDE.md, the project's design system, routed sibling skills.
  4. Sibling-skill ownership: route, do not duplicate.
  5. This skill's standards (below).
  6. General interface and platform conventions (WCAG 2.2, NN/g, Apple HIG, Material, GOV.UK), which references/rules.md cites per rule.

What it can do on your machine

Read from SKILL.md and the folder at commit cef4cfa. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Product Design loads about 2.8k tokens when it runs, and up to ~14k if it reads all its reference files. Until then it costs about 57 tokens; SKILL.md has 1,322 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~57
When it runs · the whole SKILL.md, loaded when a task matches
~2.8k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~14k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from mblode/agent-skills at commit cef4cfa, republished under its MIT licence (© mblode). 1,322 words, ~2,774 tokens.

Download SKILL.mdSave it as .claude/skills/product-design/SKILL.md (or your agent's skills folder). This skill also uses 6 other files; get the full folder from GitHub.
name
product-design
description
Specifies interaction choices, action scope, reversibility, recovery, and reachable states. Use when asked to "design the flow", "should delete be undoable", "choose the control", or "review this product decision".

Product Design

Decide what the interface should do, then route who builds and verifies it: pick the right interaction, make scope and consequence clear, cover reality beyond the happy path.

  • IS: the decision layer. From a brief, spec, mockup, intent, or existing UI: choose the interaction and control, name the object, scope, and consequence of each action, settle reversibility and the safeguard it implies, enumerate every reachable state, set resilience expectations, and require accessibility as task completion. It decides, then routes build, verification, and wording out.
  • IS NOT:
    • whether a feature deserves investment or fits the product: out of scope; state it as an open product question.
    • building or styling UI, visual direction, palettes, type: ui-design.
    • auditing the built result (rendered quality, a11y markup, keyboard, layout, performance, React or Next code-level UX with a ship verdict): ui-design Audit mode.
    • copy wording, persuasion, or AI-ism removal: ghostwriter.
    • motion, gesture physics, or deep typography: ui-animation, typography-audit.
    • whether an agentic feature earns trust: ax-audit.

Routing boundary

product-design owns action semantics, scope, reversibility, and contested state choices. ui-design builds and styles those states. ui-animation owns timing, gestures, and measured motion. A routine missing loading or error state stays with the UI build; a gesture replacing a control needs a product decision and an accessible alternative before its physics.

Request modes

Resolve one mode from the user's verb and artifact before acting, then load that mode's references. references/rules.md loads in every mode: every finding cites a rule ID from it, and you cannot conclude that no rule governs a decision without the registry in front of you.

ModeDispatch when the user asks forLoad (plus references/rules.md)
shape (default)"design the flow for", "what control here", "how should this work", "is this the right pattern", a brief with no settled UIreferences/product-judgment.md, references/surfaces.md
spec"spec the right interaction", "define the expected states", judgment applied before or during a buildreferences/product-judgment.md, references/surfaces.md, references/naming-and-copy.md; route the build to ui-design
review"review this flow for product correctness", "what's wrong with this UX decision", "is this the right interaction"references/interface-quality.md
action"what should this action affect", "should this be undoable", "do we need a confirm dialog", or reversibility is unsettledreferences/naming-and-copy.md; route final wording to ghostwriter
harden"make this resilient", "what breaks here", error, permission, offline, expiry, and destructive pathsreferences/surfaces.md, references/interface-quality.md, references/product-judgment.md

Review mode is about a flow, not an artifact. "Audit this component", "check my UI", and "design QA this page" point at built markup and belong to ui-design Audit mode. This skill's review asks whether the decisions behind a flow are right, and stops at decision altitude.

Modes chain: shape leads into spec; review leads into harden. When intent is ambiguous, use the narrowest mode the verb supports. A URL, screenshot, route, or component identifies scope; it does not authorize edits.

Whether a standard belongs in the consuming project's linter rather than in this skill is codebase-architecture's enforcement ladder. evals/evals.json never loads during a user task; it is the regression set for anyone changing this skill.

Workflow

text
Product design pass:
- [ ] Step 1: Classify the request into one mode
- [ ] Step 2: Locate authority (user constraints, project design system, AGENTS.md)
- [ ] Step 3: Load rules.md plus that mode's reference files
- [ ] Step 4: Write the internal brief (shape, spec, harden); stop and ask if job, outcome, or consequence is unfillable
- [ ] Step 5: Name object, scope, consequence, and reversibility for each action in scope (spec, action, review)
- [ ] Step 6: Enumerate reachable states and check coverage (shape, spec, harden)
- [ ] Step 7: Emit output with a rule ID or labeled coverage gap per finding or decision; route follow-on work to siblings
- [ ] Step 8: List unresolved product decisions

Steps 5 and 6 are mode-scoped because their references are: a pure action pass has no state matrix to enumerate, and a shape pass has no built actions to name yet.

Output length follows the work, not the template. A single settled decision is a short answer; drop the sections a pass did not need rather than filling them.

Decision authority

Conflict order, highest first:

  1. The user's explicit goal and constraints.
  2. Verified user and product evidence, and what the system actually does.
  3. Project-canonical guidance: AGENTS.md or CLAUDE.md, the project's design system, routed sibling skills.
  4. Sibling-skill ownership: route, do not duplicate.
  5. This skill's standards (below).
  6. General interface and platform conventions (WCAG 2.2, NN/g, Apple HIG, Material, GOV.UK), which references/rules.md cites per rule.

When a request spans authorities, name the owning skill and hand off.

Product design standards

Five pillars, each naming its rule IDs in references/rules.md and the reference that details it.

  • Right interaction. Pick the control from the choice's shape; keep options visible and reversible; prefer inline disclosure over a modal; every gesture has a control alternative; choose the smallest coherent intervention. rule/control-matches-cardinality, rule/navigation-vs-action, rule/inline-before-modal, rule/no-nested-modals, rule/gesture-has-control-alternative, rule/smallest-intervention. Detail: references/product-judgment.md.
  • Action naming and consequence. Name the object, scope, and consequence; settle reversibility first, then the pattern; an irreversible action gets review, check, or named confirmation; undo appears only when honest. rule/name-object-scope-consequence, rule/destructive-names-action, rule/destructive-proportional, rule/irreversible-action-safeguard, rule/undo-only-when-honest, rule/preserve-user-input. Detail: references/naming-and-copy.md.
  • State coverage. Design every reachable state, not just the populated one: empty states name the object and a first action; errors explain and offer recovery; timers warn before they discard. rule/cover-reachable-states, rule/empty-state-action, rule/error-states-recovery, rule/loading-stable-labels, rule/time-limit-adjustable. Detail: references/surfaces.md.
  • Resilience. Overflow, extreme data, localization and RTL, offline, and network failure are designed states; every fetch lands in one. Shares rule/cover-reachable-states: it is the same requirement pointed at adverse inputs. Detail: references/surfaces.md > Resilience. Whether the built UI renders them is ui-design Audit mode's check.
  • Accessibility as task completion. Every control has a name; the primary flow completes by keyboard with visible focus; nothing already entered is retyped; authentication allows assistance; state and consequence are understandable, not just labeled. rule/accessible-name-required, rule/keyboard-complete-flow, rule/no-custom-focus-bypass, rule/no-redundant-entry, rule/auth-allows-assistance. Detail: references/interface-quality.md. Markup and target-size checks route to ui-design Audit mode.
Show full SKILL.md (456 more words)Show less

Review output

In review and harden modes, lead with findings ordered by user impact (P0 to P3), each with location, verification status, rule ID, user consequence, and the smallest concrete fix with the skill that owns it. Keep findings at decision altitude; a line-level code or framework fix is ui-design Audit mode's output. Rubric and finding format: references/interface-quality.md > Severity rubric.

Gotchas

  • A confirmation dialog on a reversible action (archive, remove from list, unsubscribe) trains users to click through, so the one confirmation that matters, permanent delete, gets the same reflexive click. NN/g's "cry wolf" finding. Act and offer undo instead (rule/destructive-proportional).
  • An undo toast as the only recovery path: it vanishes in about five seconds, and the object is gone. Either the object stays recoverable (Trash, Archive) or the action is irreversible and needs a safeguard. Do not call it undo otherwise (rule/undo-only-when-honest, WCAG 2.2.1).
  • Swipe-to-delete or drag-to-reorder with no button or menu equivalent fails WCAG 2.5.1 and 2.5.7 outright and is undiscoverable to everyone else. Spec the alternative before routing the physics to ui-animation (rule/gesture-has-control-alternative).
  • A segmented control with three fixed-width English labels: strings under 10 characters grow 200 to 300% in translation (W3C), so "Day / Week / Month" becomes a wrapped or clipped mess in German. Decide the wrap or stack behavior in the spec.
  • Reusing the "No projects yet, Create project" empty state for a filtered-to-zero list. The user creates a duplicate because the item they searched for exists behind the filter. Three empties, three designs (references/surfaces.md).
  • A checkout or sign-up that asks again for something entered two steps earlier (billing address after shipping, email after account) fails WCAG 3.3.7 and is the step where mobile users leave (rule/no-redundant-entry).
  • Emitting a line-level fix (a prop, a hook, a className) instead of the decision. It arrives without the rendered check that would validate it, and the product decision it was supposed to carry goes unstated. Route it to ui-design Audit mode.
  • Citing a plausible-sounding rule ID that does not exist (rule/clear-labels). The citation resolves to nothing, so the finding cannot be deduped against a ui-design audit or traced to a rule. Record a coverage gap instead.
  • ui-design: visual direction and building the decided interaction in code; its Audit mode covers the built result, rendered quality and accessibility markup, with a ship verdict.
  • ghostwriter: exact wording for names, errors, and empty and loading copy, written against the copy rule IDs this skill owns in references/rules.md.
  • ui-animation: the passage between two states (timing, easing, springs, gesture physics). This skill settles whether a gesture replaces a control and what its alternative is; that skill builds the motion.
  • ax-audit: whether a built agentic feature earns trust; this skill decides what it should do first.
  • typography-audit: deep type.

© mblode, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 6 other files (references) in skills/product-design of mblode/agent-skills.

  • SKILL.md
  • evals/evals.json
  • references/interface-quality.md
  • references/naming-and-copy.md
  • references/product-judgment.md
  • references/rules.md
  • references/surfaces.md

Open the folder on GitHubat commit cef4cfa

Compare with similar skills

Product Design next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Product Design compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Product Design this skillmblode/agent-skills143—~2.8kAutomated safety check: PassMIT
Web Artifacts Builderanthropics/skills180k41 repos~769Automated safety check: PassApache-2.0
React Doctormakeplane/plane61k12 repos~657Automated safety check: PassAGPL-3.0
Impeccablebestofjs/bestofjs3.1k27 repos~2.6kAutomated safety check: PassMIT
Figma Design System Builderwarpdotdev/warp65k2 repos~4.4kAutomated safety check: PassAGPL-3.0
Web Interface Guidelines Reviewervercel-labs/openreview1.7k98 repos~308Automated safety check: PassNone

Similar skills

  • Web Artifacts Builder

    anthropics/skills

    Official

    Builds multi-component claude.ai HTML artifacts as a small React, TypeScript and Tailwind project, then bundles it into one shareable HTML file.

    180k GitHub starsUsed in 41 repos~769 tokens
    Frontend & DesignAuto-check passed
  • React Doctor

    makeplane/plane

    Scans React code for lint, accessibility, bundle size and architecture issues, reports a health score and checks that changes do not lower it.

    61k GitHub starsUsed in 12 repos~657 tokens
    Frontend & DesignAuto-check passed
  • Impeccable

    bestofjs/bestofjs

    A skill your agent uses when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a…

    3.1k GitHub starsUsed in 27 repos~2.6k tokens
    Frontend & DesignAuto-check passed
  • Builds or updates a design system in Figma from a codebase in ordered phases: discovery, variables and tokens, components, theming and documentation, with checkpoints.

    65k GitHub starsUsed in 2 repos~4.4k tokens
    Frontend & DesignAuto-check passed
  • Web Interface Guidelines Reviewer

    vercel-labs/openreview

    Official

    Review UI code for Web Interface Guidelines compliance. Use when asked to "review my UI", "check accessibility", "audit design", "review UX", or "check my…

    1.7k GitHub starsUsed in 98 repos~308 tokens
    Frontend & DesignAuto-check passed
  • Tailwindcss Development

    anonaddy/anonaddy

    Always invoke when the user's message includes 'tailwind' in any form.

    4.9k GitHub starsUsed in 10 repos~865 tokens
    Frontend & DesignAuto-check passed

More from mblode/agent-skills

All 28 skills in this repo
  • Agent Ready

    mblode/agent-skills

    Implements agent-readiness on public sites and docs from Mintlify Agent Score, AFDocs, Is Agentic, Is It Agent Ready, or url-discovery-bench reports, or from server logs of agents 404ing on guessed…

    143 GitHub stars~2.1k tokensUpdated 2 days ago
    Auto-check passed
  • Agent Skills Creator

    mblode/agent-skills

    Creates and improves portable Agent Skills with a validator, routing scenarios, and evidence-based keep, cut, merge, or retire decisions.

    143 GitHub stars~2.8k tokensUpdated 2 days ago
    Auto-check passed
  • Chat History

    mblode/agent-skills

    Recovers decisions, previous fixes, research, and what followed a prompt from past AI conversations, with source evidence.

    143 GitHub stars~1.5k tokensUpdated 2 days ago
    Auto-check passed
  • CI Speedup

    mblode/agent-skills

    Cuts the wait from push to green by measuring a pipeline's critical path from run timestamps, then splitting, sharding, trimming setup and sharing test module state, with a before/after ledger.

    143 GitHub stars~2.4k tokensUpdated 2 days ago
    Auto-check passed
  • PR Babysitter

    mblode/agent-skills

    Monitors or repairs an open GitHub PR: CI failures, conflicts, review threads, and merge readiness, reporting state changes.

    143 GitHub stars~3.4k tokensUpdated 2 days ago
    Auto-check passed
  • App Verification

    mblode/agent-skills

    Builds and maintains a repo's own verification harness (verify CLI, doctor, worktree isolation, feature map, seed data) and a reproduce-first bug handoff.

    143 GitHub starsUsed in 1 repo~2.3k tokens
    Auto-check passed

Questions about Product Design

What does Product Design do?

Specifies interaction choices, action scope, reversibility, recovery, and reachable states. Product Design is an agent skill from mblode/agent-skills. Specifies interaction choices, action scope, reversibility, recovery, and reachable states.

When should I use Product Design?

Product Design fits situations like: asked to design the flow; should delete be undoable; choose the control; review this product decision.

How do I install Product Design in Claude Code?

Run `npx skills add mblode/agent-skills --skill product-design -a claude-code`. Or copy the skill folder (skills/product-design in mblode/agent-skills) into .claude/skills/product-design in your project. Claude Code loads it when a task matches its description.

How do I install Product Design in Codex?

Run `npx skills add mblode/agent-skills --skill product-design -a codex`. Or copy the skill folder (skills/product-design in mblode/agent-skills) into .agents/skills/product-design in your project. Codex loads it when a task matches its description.

Can I use Product Design in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add mblode/agent-skills --skill product-design -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/product-design, .gemini/skills/product-design, .github/skills/product-design and .opencode/skills/product-design in your project.

What does Product Design need to run?

SKILL.md names no scripts, command-line tools or credentials: Product Design is instructions for the agent only.

Does Product Design access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Product Design safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Product Design use?

Product Design is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Product Design use?

About 2.8k tokens (SKILL.md is roughly 11k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 11k tokens, read only when the agent opens those files.

What are the alternatives to Product Design?

Skills that share tags, products or a category with Product Design: Web Artifacts Builder (anthropics/skills, 180k stars), React Doctor (makeplane/plane, 61k stars), Impeccable (bestofjs/bestofjs, 3.1k stars) and Figma Design System Builder (warpdotdev/warp, 65k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Product Design?

mblode (a GitHub user) maintains it in mblode/agent-skills, which has 143 GitHub stars. The repository holds 28 skills in this directory. The repository was last updated on October 6, 2026.

Source: mblode/agent-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.