Design Judge Skills — Setup Guide

Repo: SeanJ1ang/design-judge-skills Stars: 25 · Forks: 6 · License: Apache 2.0 Language: Python · Skills: 5 + 1 shared support package Agent Support: Hermes Agent, Claude Code, Codex, OpenClaw, OpenCode

Overview

Design Judge Skills decomposes the end-to-end design award submission process into five modular, evidence-driven agent skills. Each skill operates independently with well-defined boundaries — you can use one, several, or all five as your workflow demands.

The skills are configured for 11 major design awards: iF DESIGN AWARD, iF DESIGN STUDENT AWARD, Red Dot Product Design, Red Dot Design Concept, IDEA, DIA, K-Design, GOOD DESIGN AWARD Japan, Core77, James Dyson, and EPDA. Award rules, deadlines, fees, eligibility, and categories are verified against official pages at runtime.

The 5 Skills

1. design-award-search — Find Winning Precedents

Retrieve and verify winning entries in the same category from official award archives.

Use when: You need examples of what won in your category, or want to understand what juries reward.

Trigger phrase: "Find winning cases for [product type] in [award/category]"

Example prompt:

Use design-award-search to find verified winning entries for a rehabilitation training device
in the iF DESIGN AWARD product category. Verify each case against the official iF archive.

2. design-evaluation — Score Design Quality

Evaluate design quality and presentation expression using evidence-based rubrics with transparent scoring.

Use when: You need an objective assessment of a design before submitting it to awards.

Trigger phrase: "Evaluate this design for [award] submission"

Example prompt:

Use design-evaluation to assess the attached design. Maturity level is "student concept."
Output design quality score, presentation expression score, evidence confidence, and critical issues.

Output structure: - Design quality score (with rubric alignment) - Presentation expression score - Evidence confidence (what's verified vs. assumed) - Critical issues (blockers before submission)

3. design-award-match — Compare Award Fit

Compare awards, tracks, and submission categories against project characteristics to find the best fit.

Use when: You have multiple award options and need to decide where to submit.

Trigger phrase: "Compare award fit for [project]"

Example prompt:

Use design-award-match to compare iF Student, Red Dot Design Concept, DIA, Core77,
and James Dyson for this project. Output fit scores with explanations for each.

4. design-information-prep — Prepare Submission Text

Prepare award submission text constrained to source materials, with field-by-field drafting and word-count validation.

Use when: You have design materials and need to write submission copy.

Trigger phrase: "Prepare submission text for [award]"

Example prompt:

Use design-information-prep to prepare IDEA award submission text from the attached materials.
First list missing facts, then output English draft per field with word count verification.

5. design-submission-check — Final Readiness

Check a submission package against the current official rules of the target award. Outputs a go / conditional go / no-go verdict.

Use when: Your submission package is complete and needs a final compliance review.

Trigger phrase: "Check submission readiness for [award]"

Example prompt:

Use design-submission-check to verify this submission package against the current iF DESIGN AWARD
official rules. Output go / conditional go / no-go with itemized issues.

Shared Support Package

design-judge-shared is a shared support package used by the other five skills. It's automatically included when any skill is installed. You don't invoke it directly.

Design Principles (from the repo)

These principles govern all five skills. They're worth understanding before you rely on the output:

1. Primary sources first. Award rules and cases verified against official pages. Search summaries and third-party pages are for discovery only.
2. Facts vs. inference separated. User materials, model inferences, and items requiring user confirmation are labeled distinctly.
3. Transparent scoring. Fit and evaluation scores aid decisions but don't simulate judging panels or predict outcomes.
4. Single-responsibility modules. Search, evaluation, matching, text prep, and submission check don't cross boundaries.
5. No fake judging. Skills align with public criteria but don't simulate undisclosed judge preferences.

Installation

Option 1: npx skills (Recommended — works with any agent)

Install individual skills:

npx skills add SeanJ1ang/design-judge-skills --skill design-award-search -g -y
npx skills add SeanJ1ang/design-judge-skills --skill design-evaluation -g -y
npx skills add SeanJ1ang/design-judge-skills --skill design-award-match -g -y
npx skills add SeanJ1ang/design-judge-skills --skill design-information-prep -g -y
npx skills add SeanJ1ang/design-judge-skills --skill design-submission-check -g -y

The -g flag installs at global/user scope. The -y flag skips confirmation prompts. Remove -y for interactive installation.

Option 2: Hermes Agent — Manual Skill Installation

Copy the skill directories to your Hermes profile:

# Clone the repo
git clone https://github.com/SeanJ1ang/design-judge-skills.git /tmp/design-judge-skills

# Copy skills to Hermes profile
cp -r /tmp/design-judge-skills/skills/* ~/.hermes/profiles/corpusiq/skills/

# Clean up
rm -rf /tmp/design-judge-skills

Then reload skills in your Hermes session or restart.

Option 3: Claude Code

# In Claude Code session
/install-skill SeanJ1ang/design-judge-skills/design-award-search

Replace design-award-search with any of the five skill names.

Option 4: Codex

codex skills install SeanJ1ang/design-judge-skills --skill design-award-search

Verification

After installation, verify skills are loaded:

Hermes Agent:

ls ~/.hermes/profiles/corpusiq/skills/design-*/

npx skills:

npx skills list | grep design-

Quick test: Describe a design and ask the agent which award to target. The agent should invoke design-award-match:

I have a wearable health monitor for elderly users. It tracks heart rate, fall detection,
and medication reminders. Which design awards should I target?

Workflow

The five skills form a pipeline but can be used independently:

Design Materials
  ├─ Find winning precedents ──────> design-award-search
  ├─ Evaluate design quality ──────> design-evaluation
  └─ Match awards & categories ───> design-award-match
                                         │
                                         v
                                design-information-prep
                                         │
                                         v
                                design-submission-check

A complete submission flow typically starts with evaluation or matching, moves to text preparation once the target award is selected, and ends with a submission check.

Supported Awards

Award	Category Coverage	Source Verification
iF DESIGN AWARD	Product, Communication, Packaging, Service Design, UX, Professional Concept	Official iF archive
iF DESIGN STUDENT AWARD	Student concepts across all categories	Official iF archive
Red Dot Product Design	51 product categories	Official Red Dot archive
Red Dot Design Concept	Concepts, prototypes, student work	Official Red Dot archive
IDEA	Professional + Student	IDSA official site
DIA (Design Intelligence Award)	Product, Concept, Digital	DIA official site
K-Design Award	Product, Space/Architecture, Communication, Service, Fashion	K-Design official site
GOOD DESIGN AWARD Japan	All categories	G-Mark official archive
Core77 Design Awards	23 categories	Core77 official site
James Dyson Award	Student design + engineering	James Dyson Foundation
EPDA (European Product Design Award)	Product, Packaging, Interface	EPDA official site

Award rules, deadlines, and fees are verified at runtime. Do not cache these — they change annually.

Limitations

• Runtime verification required. Award rules, deadlines, and fees must be re-verified against official pages at runtime. The skills don't cache this data.
• Language. Primary documentation is Chinese with English translation. Skills accept materials in either language.
• Not a judging simulation. Evaluation scores are heuristic aids, not predictions of jury outcomes.
• 11 awards configured. Additional awards require adding configuration to the shared support package.

Security

• The skills read design materials you provide and search official award websites. No data is sent to third-party services.
• All source code is in Python and the shared support package. Review skills/design-judge-shared/ before using with confidential designs.
• License: Apache 2.0 — safe for commercial use.

Troubleshooting

"Skill not found" after npx install

The npx skills CLI caches skills. Try:

npx skills list --refresh | grep design-

Award rules mismatch

Awards update annually. If the skill references outdated rules, specify the year in your prompt:

Use design-submission-check for iF DESIGN AWARD 2027 rules.

Language issues

The repo is primarily Chinese with English README. If outputs appear in Chinese, specify language in your prompt:

Output all results in English.

---

Setup guide by CorpusIQ. Skill repo by SeanJ1ang. Licensed Apache 2.0.