Memi Design Tooling

SkillDocs & knowledge

Use when a task spans interface understanding, design-system memory, UI audits, design CI, Figma, shadcn or Tailwind code generation, research, or agent design workflows and needs the correct Memi capability selected.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Memi Design Tooling skill

What this skill tells your AI

The instructions your AI receives, as published by memi-design/memi in skills/memoire-design-tooling/SKILL.md and read by ahel’s review.

Memi gives coding agents repository-specific interface evidence before they edit UI. Start with the smallest workflow that answers the task; Figma, global installation, and a daemon are optional.

Choose A Workflow

  • Before reviewing or changing frontend UI: use audit-frontend-design.
  • Before building from an existing product system: use remember-design-system.
  • When adding deterministic pull-request gates: use enforce-design-ci.
  • For native SwiftUI, SwiftData, App Intents, or Apple-platform verification: use build-swiftui-interface.
  • For Figma, research, scaffolding, registry publishing, or multi-agent work: continue below.

Install one focused skill directly:

npx skills add memi-design/memi --skill audit-frontend-design

This workflow targets published beta 2.8.0-beta.2. Use npx -y @memi-design/cli@2.8.0-beta.2 --version to verify the exact package; a local memi must report that same version before these recipes run. The independent stable compatibility baseline is 2.7.9. See current release state and known limitations. npm availability alone does not establish native-binary success.

Compact Preflight

memi agent brief . --frontend --intent "<interface task>" --max-bytes 16384 --json

Read cited source files when the bounded brief lacks necessary evidence. Optional --design-evidence design/selected-node.json accepts normalized Figma/Paper data supplied by the harness; missing mappings and verification stay explicit.

Local Checks And MCP

memi diagnose . --json --no-write --fail-on none
memi diagnose . --receipt-only --fail-on none
memi --profile locked mcp start --no-figma

Locked MCP exposes four read tools: prepare_frontend_brief, prepare_design_agent_brief, prepare_apple_design_brief, and diagnose_app_quality. The frontend tool supplies actual repository evidence. Many legacy CLI and write tools remain unavailable; capability grants do not unlock deferred command paths. Keep connector calls, project edits, and browser execution in the harness's reviewed workflow.

Reuse local components and tokens according to the consumer project's conventions. Source-bearing JSON is working context; receipt-only is a separate metadata output. See docs/FRONTEND_WORKFLOW.md for the beta evidence schema and actual verification flow.

Evidence Contract

  1. Read local instructions and existing product-system files first.
  2. Collect the minimum evidence that can change the implementation.
  3. Cite file:line findings and existing components or tokens.
  4. Make scoped edits.
  5. Re-run the same deterministic checks.
  6. Report commands, artifacts, files changed, and remaining assumptions.

Do not claim visual correctness from source checks alone. When rendered behavior matters, verify the actual route at desktop and mobile viewports.

Signals

GitHub stars
43
Forks
5
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
memoire-design-tooling
Source
github.com/memi-design/memi