reference-interpreter
Skill194 starsChanged 7 months agoMITView on GitHub
Triggers when user shares a screenshot, image, URL, or design description — or says 'analyze this', 'make a brief', 'interpret this reference'. Outputs a structured Design Brief mapping visual intent to design system tokens. Waits for 'confirmed' before designing.
In this repository3 other files it gives its agents.
- component-rulesSkill
- figma-preflightSkill
- figma-style-bindingSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 38
If no token exists, flag it:
- Very specific: line 88
Do NOT call use_figma or place any nodes until user types confirmed.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Reference Interpreter
Analyze a reference (screenshot, URL, or description) and produce a Design Brief that maps visual observations to design system tokens. Output the Brief, then stop and wait for user confirmation before building anything.
Prerequisite: figma-preflight should have run so the Token Map and Style Registry are available.
Phase 1 — Analyze
Examine the reference across these dimensions:
- Layout — structure, columns, section heights, grid width, alignment
- Typography — heading/body hierarchy, weight contrast, tracking
- Color — dark/light sections, accent usage, neutral dominance
- Spacing — generous vs compact, section padding, internal gap
- Visual Anchor — what draws the eye: large type, hero image, illustration
- Components — what UI elements are visible: cards, buttons, forms, nav
Phase 2 — Map to Design System
For each observation, identify the specific Token or Style from the session's Token Map:
"Large dark headline" → Text Style: heading/h1 · Color: text/primary
"Neutral section bg" → Variable: background/surface-2
"Tight card spacing" → Gap: gap/xs · Padding: padding/smIf no token exists, flag it:
⚠️ Gap: [observation] — no matching token. Options: (a) nearest match: [name], (b) add token first.Phase 3 — Output the Design Brief
## Design Brief
**Reference**: [source]
**Section**: [what this covers]
**Aesthetic**: [3-5 keywords]
### Layout
[structure, height, alignment]
### Typography
- Heading: [Style name] — [why]
- Body: [Style name]
### Colors
- Background: [Variable] — [context]
- Text: [Variable]
- Accent: [Variable] — [used for]
### Spacing
- Section padding: [Variable]
- Internal gap: [Variable]
### Components needed
- [Name]: [from library? source]
### Gaps
- [Gap description] — awaiting decision
- (none) [if all mapped]Phase 4 — Wait
Output exactly:
Brief complete. Type `confirmed` to begin building, or tell me what to adjust.Do NOT call use_figma or place any nodes until user types confirmed.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

