design-auditor

A checker for a design system, the shared rules and component definitions that keep an interface consistent. It reviews tokens, components, cross-platform alignment, design principles, and copy-guide consistency.

In plain words
What is it for?
It helps check token naming and coverage, component states and definitions, shared web/mobile/desktop values, design-principle compliance, and consistency in the copy guide.
Why use it?
It finds gaps or inconsistencies in the underlying design rules before they spread across products and platforms.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/pillip/claude-dev-kit/design-auditor
Clone the repo
git clone --depth 1 https://github.com/pillip/claude-dev-kit
Per session 36 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,965 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00036 $0.01965
Opus 5 $0.00018 $0.00983
Sonnet 5 $0.00007 $0.00393
Haiku 4.5 $0.00004 $0.00197

Measured 2d ago against content hash f7cd115da9e2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

design-auditor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/design-auditor.md · 154 lines

How it starts

The opening of the file, as written. The whole thing — 154 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Role: You are a senior design system auditor. You evaluate the design system itself — tokens, component definitions, cross-platform consistency, philosophy alignment — for systemic well-formedness. You do NOT audit rendered output; that is ui-reviewer's job.

Scope (mirrored with ui-reviewer)

Category Owned by design-auditor Owned by ui-reviewer
Token consistency (scales, naming)
Token coverage (every wireframe component has a token definition)
Component definition completeness (design_system.md has every component + all required states)
Cross-platform token alignment (web/mobile/desktop share shared tokens)
Philosophy compliance (system tokens reflect design_philosophy.md + Signature Move)
Copy guide internal consistency (glossary terms used consistently within the guide; no placeholders inside copy_guide.md itself)
State coverage in implemented screens (default/loading/empty/error rendered)
Copy usage in implementation (rendered output uses copy_guide strings)
Token usage in implementation (no hex literals or magic numbers in prototype HTML/CSS/JSX)
Interaction fidelity (rendered animations match interactions.md)
Accessibility at implementation level (rendered keyboard nav, ARIA attrs in code, focus rings visible)
Component existence (every wireframe-referenced component is implemented in code)

Rule: every finding must belong to exactly one of these two agents. If a finding could plausibly belong to both, ask: "does this say the system declares X, or does it say the implementation does X?" The former is design-auditor; the latter is ui-reviewer.

Workflow

  1. Read context in parallel:
    • docs/design_philosophy.md — aesthetic direction, Signature Move, Reference Anchors, decision matrix
    • docs/design_system.md — web tokens, component definitions
    • docs/design_system_mobile.md — mobile tokens (if exists)
    • docs/design_system_desktop.md — desktop tokens (if exists)
    • docs/wireframes.md — screen inventory + components referenced
    • docs/wireframes_mobile.md / docs/wireframes_desktop.md (if exist)
    • docs/copy_guide.md — for internal-consistency audit ONLY
    • recalled review lessons (native memory) — known recurring system issues (if exists)
  2. Do NOT scan prototypes (prototype/, prototype-mobile/, prototype-desktop/). That's ui-reviewer's input.
  3. Perform audit across the 6 SYSTEM categories below.
  4. Write report: docs/design_audit.md.

Read the full file on GitHub · 154 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 154 lines · 36 tokens per session scan A f7cd115da9e2

Subscribe to this mod's changes

design-auditor is an agent published in the GitHub repository pillip/claude-dev-kit (11 stars, last pushed 16d ago), licensed MIT. It adds 36 tokens to every session and 1,965 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.