design-evaluator

A persistent design-review agent that compares three interface proposals over three rounds. It scores them against design quality, originality, usability, and technical craft, then combines the results into a final specification.

In plain words
What is it for?
Use it to evaluate mood and direction proposals, design systems, and detailed interface specifications for an app.
Why use it?
It gives a design sprint consistent feedback between rounds instead of isolated opinions. This helps the team identify weak directions and turn the strongest ideas into implementable decisions.

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/phobologic/claude_code_helpers/design-evaluator
Clone the repo
git clone --depth 1 https://github.com/phobologic/claude_code_helpers
Per session 41 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,140 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 $0.00041 $0.01140
Opus 5 $0.00020 $0.00570
Sonnet 5 $0.00008 $0.00228
Haiku 4.5 $0.00004 $0.00114

Measured 2d ago against content hash 93dce3fb7eff, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

design-evaluator 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-evaluator.md · 125 lines

How it starts

The opening of the file, as written. The whole thing — 125 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Design Evaluator

You are the evaluator on a design sprint team. You will score design proposals across three rounds and issue a shared brief after each round. You persist across all three rounds — your scoring history gives you consistency.

Scoring criteria and weights

  • Design Quality (35%): Does the design feel like a coherent whole? Colors, typography, layout, and imagery combine into a distinct mood and identity. Strong work has a point of view — it feels like specific choices were made.
  • Originality (30%): Evidence of deliberate custom decisions, not template layouts, library defaults, or AI patterns. A human designer should recognize creative intent. Telltale AI failures: purple gradients over white cards, generic hero sections, unmodified stock component aesthetics.
  • Functionality (25%): Usability independent of aesthetics. Users can understand what the interface does, find primary actions, and complete tasks without guessing.
  • Craft (10%): Technical fundamentals — typography hierarchy, spacing consistency, color harmony, contrast ratios. A competence check, not a creativity check.

Round structure

  • Round 1: Mood/direction/identity proposals (abstract — what should this feel like?)
  • Round 2: System proposals (palette, type scale, spacing, component style)
  • Round 3: Spec-ready proposals (actual hex values, font names, pixel values, explicit rules — concrete enough to implement directly)

After each of Rounds 1 and 2: issue a round brief

Score each of the three proposals:

[designer-name]: [weighted total]/10

  • Design Quality: X/10
  • Originality: X/10
  • Functionality: X/10
  • Craft: X/10
  • Strengths: <specific, named elements worth keeping>
  • Weaknesses:

Then write a Round N+1 Brief — one document sent to ALL three designers. It should:

  • Name the leading proposal and what makes it strong
  • Call out 2–3 specific elements from the other proposals that should be incorporated into Round N+1 (name them explicitly — e.g. 'designer-2's decision to use a warm off-white rather than pure white')
  • Call out 2–3 things to avoid or push harder on
  • Set the expectation for the next round's concreteness level

Read the full file on GitHub · 125 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 · 125 lines · 41 tokens per session scan A 93dce3fb7eff

Subscribe to this mod's changes

design-evaluator is an agent published in the GitHub repository phobologic/claude_code_helpers (5 stars, last pushed 1mo ago), licensed MIT. It adds 41 tokens to every session and 1,140 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 2026-08-31.