agentleFS
thehelpfultipperthehelpfultipper/wp-agent-mcp/.cursor/skills/wp-agent-toolkit-design-bridge/SKILL.md

wp-agent-toolkit-design-bridge

Skill1 starsChanged 33 days agoMITView on GitHub

Bridge Figma or screenshot design intent into WordPress repo conventions via WP Agent MCP. Use when the user asks to implement, build, or match a design in WordPress.

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. About 0 are common, 3 are distinctive, and 11 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 11

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Call get_design_bridge_readiness (WP Agent MCP) first.

    line 20
  2. Very specific:

    If readyForBridge: false, explain blockers[] and remediation before capture work.

    line 21
  3. Very specific:

    Before Figma capture, verify Figma MCP via dynamic tool discovery (GetDynamicTools pattern figma).

    line 26
  4. Distinctive:

    If missing and user gave a Figma URL: explain how to enable Figma MCP in Cursor Settings → MCP, then offer screenshot fallback (no Figma …

    line 27
  5. Distinctive:

    Do not spawn parallel subagents for bridge steps in v1.

    line 32
  6. Very specific:

    Do not ask the user for themeSlug, blogId, or other tool fields.

    line 56
  7. Very specific:

    If the bridge returns requiresThemeSelection: true, ask the user once in natural language which theme to use (offer themeCandidates), then …

    line 57
  8. Very specific:

    Call bridge_design_to_implementation with visualIntent (or figmaCapture) and componentHint from the user's request.

    line 61
  9. Distinctive:

    Use the returned implementation spec:

    line 63
  10. Very specific:

    Read cited files via read_project_file

    line 72
  11. Very specific:

    Prefer reuse/extend per componentPlan

    line 73
  12. Very specific:

    If Playwright MCP available: run verificationChecks

    line 78
  13. Very specific:

    After one get_design_context call:

    line 91
  14. Very specific:

    Add components[].label for each component/frame name (include block slugs such as tht/card-hero when visible).

    line 94

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Design Implementation Bridge

When to use

Route here when the user:

  • Shares a Figma link or node and wants it implemented in WordPress
  • Attaches a screenshot/mock and asks to match the project design system
  • Says "implement this frame/section/hero" with design context

Users stay in natural language. Never require tool names in user prompts.

Prerequisites (transparent checks)

  1. Call get_design_bridge_readiness (WP Agent MCP) first.
    • If readyForBridge: false, explain blockers[] and remediation before capture work.
    • Surface companionMcps.figma.setupHint when the user provided a Figma URL.
    • Surface companionMcps.playwright.setupHint only for post-implementation verification.
  2. Figma MCP availability (agent-side only)
    • Before Figma capture, verify Figma MCP via dynamic tool discovery (GetDynamicTools pattern figma).
    • If missing and user gave a Figma URL: explain how to enable Figma MCP in Cursor Settings → MCP, then offer screenshot fallback (no Figma MCP required).

Workflow (linear, quota-efficient)

  1. Readiness → capture → one interpret pass → bridge → implement → verify
  2. Do not spawn parallel subagents for bridge steps in v1.

Capture

InputAction
Figma URLOne get_design_context call (load figma-design-to-code skill first)
Screenshot/imageInterpret image directly; skip Figma MCP

Interpret (agent, one pass)

Convert capture into semantic visualIntent:

  • Layout structure + region roles
  • Color roles (background, text, accent, …) — not raw hex as the primary signal
  • Typography roles (heading, body, …)
  • Spacing rhythm (tight / normal / loose, sectionPadding)
  • Component labels/kinds

Raw Figma values may appear under raw hints only. Never send raw-only payloads (bridge returns BRIDGE_INTENT_TOO_RAW).

Theme target

  • Pass componentHint with a natural-language excerpt from the user's request (component description and optional theme mention, e.g. "card hero in our tht theme").
  • Do not ask the user for themeSlug, blogId, or other tool fields.
  • If the bridge returns requiresThemeSelection: true, ask the user once in natural language which theme to use (offer themeCandidates), then retry with the chosen theme reflected in componentHint.

Bridge (single MCP call)

Call bridge_design_to_implementation with visualIntent (or figmaCapture) and componentHint from the user's request.

Use the returned implementation spec:

  • tokenReconciliation → theme preset CSS vars
  • componentPlan.audit → reuse / extend / create
  • implementationGuidance.avoidRawValues → never paste Figma literals into theme files
  • verificationChecks → Playwright or manual checklist

Implement

  • Read cited files via read_project_file
  • Prefer reuse/extend per componentPlan
  • Apply changes via apply_patch when appropriate

Verify

  • If Playwright MCP available: run verificationChecks
  • Otherwise: present manual checks from the spec

Anti-patterns

  • Pasting Figma hex/spacing/font values directly into SCSS/PHP/theme.json
  • Calling get_theme_profile + audit_component separately when bridge can bundle reconciliation
  • Running search_project_files, terminal grep, or rg before the bridge when implementing from Figma/screenshot (bridge bundles audit + token reconciliation)
  • Multiple Figma MCP reads for the same frame
  • Parallel Task subagents for token matching (bridge is deterministic in WP Agent)

Figma → visualIntent handoff

After one get_design_context call:

  1. Set visualIntent.layout.structure from frame layout summary.
  2. Add components[].label for each component/frame name (include block slugs such as tht/card-hero when visible).
  3. Map colors/typography/spacing to roles; keep literals under *.raw only.
  4. Or pass the capture object as figmaCapture and let the bridge derive a first-pass intent.

Include known block slugs in components[].label — layout text alone often misses indexed block matches.

Example user prompts

  • Implement this hero section in our WordPress theme + Figma link
  • [screenshot] Match this layout using our design system for the tht theme
  • Turn this Figma frame into a block pattern (ask for theme if missing)

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.