Design — 1 of 20 skills in the Architect & BTP pack
Compatible with Claude, ChatGPT, Gemini, and Claude Code.
Design-Critic
Critiques an architectural project on what actually matters.
Free, no account, no email — the command is all you need.
Already have it?
Nothing to install if it's already active
If this skill is already added in your Claude, there's no need to install it again. Just ask for what you want to do.
Not sure? Ask Claude: "which skills do you have available?"
“Use the design-critic skill to critique this architecture project (plan, section, elevation or render) on what actually matters: …”
1 · The simplest way
Copy and paste into any AI chat
This button copies the whole skill. Paste it into Claude, ChatGPT, or Gemini, then describe what you need. No installation required.
In a chat, Word or Excel files can't always be generated: you'll receive the document as structured text instead, and Claude will flag it.
2 · Command line
/design-critic installs in 3 steps
For Claude Code. The skill installs on your computer and stays available across all your projects.
Prerequisites: Claude Code + Node.js
The skill runs inside Claude Code, Anthropic's AI assistant installed on your computer. Node.js is the engine that runs the install command. Both are free.
Open PowerShell
Click Start, type "PowerShell", then press Enter.
Paste the command
Right-click inside the PowerShell window (or press Ctrl + V) to paste the command below, then press Enter. Wait a few seconds.
npx educasium --skill design-critic --globalWhat you should see:
npx educasium --skill design-critic --global
Fetching https://raw.githubusercontent.com/Chappygo-OS/claude-meta-skills/main/dist/design-critic.zip
…
Installed "design-critic". Restart Claude Code to pick it up.
Restart Claude Code
In Claude Code, type /restart (or quit and relaunch Claude Code). The skill now appears in the command list.
Don't use a terminal?
Using Claude.ai in your browser?
The assisted install works through a .zip file imported into Claude.ai. It takes longer and asks for your email.
What the skill does
View the source codeIn detail.
Hand it a plan, section, elevation, or render. Design-Critic flags 2 to 4 specific issues, each tied to a reading axis (concept consistency, scale and proportion, circulation and program, materiality and detail, context, or feasibility) and backed by something visible in the project — never a generic critique.
When to use it
Before a jury review, a client handoff, or a permit submission: when you want an outside read that won't settle for "interesting".
Example usage
/design-critic Critique this ground floor plan and section...In depth
How this skill really works
Design-Critic produces a structured, evidence-based critique of an architecture project — a plan, section, elevation, render, or massing study. Rather than grading every aspect of the project uniformly, it flags 2 to 4 specific issues, each tied to one of the six reading axes documented in references/six-axes.md (concept consistency, scale and proportion, circulation and program, materiality and detail, context, feasibility) and backed by something actually visible in the material provided — never a generic impression like "this doesn't work."
The skill judges the project against its own stated intentions: if a concept note, program, or client brief is supplied, the critique is measured against what the project claims to be doing, not a generic ideal. It never turns a critique into a design alternative — it diagnoses, it doesn't redesign.
Three tones are available to calibrate the output's severity: mentor (default, direct but warm), harsh (leads with the strongest objection, for jury prep or a tough client conversation), and encouraging (leads with what already works, for early-stage concept or schematic work). It's built for the architect preparing a jury review, a client handoff, or a permit submission who wants an outside read that won't settle for "interesting."
When to use it
- Before a jury review, a client handoff, or a permit submission, when a critical outside read is needed.
- When a plan, section, elevation, render, or massing study already exists and is ready to be assessed.
- When a concept note, program, or client brief is available to measure the project against its own stated intentions.
- When the user wants the harsh version ahead of a jury presentation the next day.
- When the work is still at concept or schematic stage and an encouraging, development-oriented tone fits better.
- When only an isolated exterior render is available, with no plans — the skill can critique what's visible while flagging the analysis's limits.
When not to use it
- Not for checking building-code or zoning compliance (dimensions, egress, setbacks) — pair with a compliance-focused skill for that.
- Not for a detailed cost or constructability review — the feasibility axis doesn't produce figures, it only questions plausibility.
- Not for generating new design options or a variant — the skill diagnoses, it doesn't redesign.
- Not from a bare description with no image or drawing — the skill asks once for the missing material rather than guessing a design from adjectives.
The method
The six reading axes
references/six-axes.md documents six reading axes, each with its own definition, what "good" looks like, and an example critique pitched at the right level of specificity. The skill selects only 2 to 4 per critique — the ones where a problem is actually visible — rather than grading all six every time.
- 1. Concept consistency
- Checks whether the stated idea actually shows up in form, material, and detail, or dilutes between concept and drawings. A reader who never saw the concept statement should be able to reconstruct it from the drawings alone.
- 2. Scale & proportion
- Assesses sizing relative to the human body, the neighboring built context, and the program the space holds. A win looks like ceiling heights and room dimensions that genuinely match what the spaces are for and what sits next to them.
- 3. Circulation & program
- Examines the logic of movement, adjacency between spaces, and where public and private actually meet. A win looks like the shortest path between two spaces that should be close also being the path people are meant to take.
- 4. Material & detail
- Checks whether the material choice supports the stated intent, or works against it once built. A win looks like the detail at the junction between two materials resolving the same idea the materials were chosen to express.
- 5. Context
- Questions the design's response to site, climate, and the regulatory or cultural conditions of the place. A win looks like a stranger being able to tell which street the building sits on from the building alone.
- 6. Innovation vs. feasibility
- Questions whether the project's ambition stays buildable within the budget and schedule it implies, even when neither is formally stated. A win looks like the project's most ambitious move also being the one with the clearest construction sequence.
Steps
Gather context
Read everything provided (drawings, images, renders) and note the project's own stated intent (concept, program, brief) if one exists. If only a vague description arrives with no image or drawing, ask once for the missing visual material rather than guessing.
Apply the six-axis framework
Scan all six axes without forcing a finding on each one. Select the 2 to 4 axes where a problem is actually visible — silence on the others is a valid, honest outcome, not a lack of thoroughness.
Choose a tone
Default to the mentor tone (direct but warm). Switch to harsh when the user flags jury prep or a tough client conversation; to encouraging for work still at concept or schematic stage.
Write the critique
Structure as a one-line synthesis, 2 to 4 findings in the format [axis] → [what is visible] → [why it matters] → [what to test next], and one next-study direction — never a redesign.
Report
Close with the top 1-2 issues restated in a single line each, and the suggested next-study direction. Don't create a file unless the user only wanted the critique inline.
Rules
- Always tie each finding to one of the six axes and to something actually visible in the material provided — never a generic impression.
- Cap the critique at 2 to 4 findings; silence on an axis with no visible issue is a valid outcome.
- Judge the project against its own stated intentions (concept, program, brief) — never against a generic standard it never claimed.
- Ask once for missing visual material if only a vague description is given, rather than guessing a design from adjectives.
- Only create an output file if the user explicitly asks for one — otherwise respond inline.
Mistakes to avoid
- Forcing a finding on all six axes instead of keeping only the 2 to 4 with a real signal.
- Critiquing the project against a standard it never claimed (e.g., a sustainability critique on a project that never set an environmental target).
- Proposing a design alternative or redrawing on the client's behalf instead of sticking to diagnosis.
- Guessing a design from an adjective-only description without ever asking for the missing drawing or image.
- Keeping the default mentor tone when the user explicitly asked for the harsh version ahead of a jury.
What this skill doesn't do
- Does not generate alternative design options, massing studies, or redraw anything.
- Does not check building-code, zoning, or constructability compliance — pair with a compliance-focused skill for that.
- Does not grade the project against a standard it never claimed (no sustainability critique without a stated sustainability target).
- Does not infer a design from a vague description alone — asks once for the missing image or drawing, then works with what it's given.
- Does not cost or detail constructability — the feasibility axis questions plausibility, it doesn't produce a costing.
Skill contents — 3 fichiers · 10.9 KB(SKILL.md, references/six-axes.md, references/tone-guide.md)
Download for free