32 deep knowledge files from 87 to 5,585 lines of real expertise each. Install them to extend what Mémoire can do. They activate automatically based on intent classification. In v0.5, Notes feed into the intelligence layer: AI vision analysis, WCAG enforcement, research-to-design mapping, and Pretext text measurement all consume Note context when agents execute.
Design craft, systems thinking, Figma plugin expertise, library building, component architecture, mobile mastery
RESEARCH
User research, competitive analysis, data synthesis, usability testing, insight extraction
CONNECT
Figma sync, Linear, Notion integrations, self-improvement, API connectors, workflow automation
GENERATE
Specialized codegen for React Native, Vue, Flutter, and Figma-to-code implementation
AVAILABLE NOTES
47 notes
CRAFTFIGMA MCP
5,585
Figma Library Builder
Complete 5-phase workflow for building Figma component libraries from code. Includes token creation, component scaffolding, documentation pages, QA validation, and 9 ready-to-use Plugin API scripts.
DiscoveryToken CreationComponent Scaffolding+2
figma-library-builder
Complete 5-phase workflow for building Figma component libraries from code. Includes token creation, component scaffolding, documentation pages, QA validation, and 9 ready-to-use Plugin API scripts.
Create variable collections with the correct architecture (Primitives + Semantic split). Create primitive variables with raw values, then semantic variables that alias primitives. Set scopes on every variable (never use ALL_SCOPES). Set code syntax on every variable.
CRAFTFIGMA MCP
3,286
Figma Plugin API
Comprehensive Plugin API reference with 14 critical gotchas, common patterns, variable and component essentials, text and effect style handling, and error recovery strategies.
API GotchasVariable PatternsComponent Patterns+2
figma-plugin-api
Comprehensive Plugin API reference with 14 critical gotchas, common patterns, variable and component essentials, text and effect style handling, and error recovery strategies.
In UI: staggered children. When a panel slides in, its contents arrive 50-80ms later. A card’s shadow settles after the card itself. Elements within a group resolve at slightly different times, creating organic layered motion.
Touch Target SystemsSafe Area HandlingPlatform ConventionsGesture DesignResponsive Patterns
memi notes install mobile-craft
CONTENT PREVIEW
The thumb zone model is based on Steven Hoober’s research on one-handed phone use. For a right-handed user holding a ~6.1" phone, the bottom 20-30% of the screen is the easy reach zone for primary CTAs, while the top 20% is hard to reach without shifting grip.
CRAFT
1,411
Design Systems
Token architecture, component API design, multi-brand theming, versioning strategy, migration patterns, and cross-platform parity. Deep knowledge for building and maintaining design systems at scale.
Token ArchitectureComponent API DesignMulti-Brand Theming+2
design-systems
Token architecture, component API design, multi-brand theming, versioning strategy, migration patterns, and cross-platform parity. Deep knowledge for building and maintaining design systems at scale.
ACTIVATES ONcomponent-creationFREEDOMhigh
ALL SKILLS
Token ArchitectureComponent API DesignMulti-Brand ThemingVersion StrategyCross-Platform Parity
memi notes install design-systems
CONTENT PREVIEW
A button never references a raw scale token directly. It references --button-bg, which resolves through --color-primary to the active brand alias. Changing the brand color only requires updating the alias tier.
CRAFTFIGMA MCP
482
Figma Design Systems
Working with Design Systems in Figma. Covers components (4 property types), variables (collections, modes, aliasing, scoping), text styles, and effect styles with creation and usage patterns.
ComponentsVariablesText Styles+2
figma-design-systems
Working with Design Systems in Figma. Covers components (4 property types), variables (collections, modes, aliasing, scoping), text styles, and effect styles with creation and usage patterns.
Variants are Figma’s reusable design entities. Layer naming encodes combinations: Variant=Primary, Size=Small, State=Disabled. Variants multiply combinatorially, so define only the axes you need.
CRAFT
325
Token Architecture
Multi-tier design token systems with semantic naming, theme switching, CSS custom properties, Tailwind mapping, and Figma variable sync. Build tokens that scale across platforms.
Semantic NamingTheme SwitchingCSS Variables+2
token-architecture
Multi-tier design token systems with semantic naming, theme switching, CSS custom properties, Tailwind mapping, and Figma variable sync. Build tokens that scale across platforms.
A button never references a raw scale token directly. It references --button-bg, which resolves through --color-primary to the active brand alias. Changing the brand color only requires updating the alias tier.
CRAFTFIGMA MCP
222
Figma DS Rules
Generate design system rules that encode codebase conventions for AI agents. Produces CLAUDE.md, AGENTS.md, or .cursor/rules files that guide consistent component generation and token usage.
Rule GenerationToken MappingAgent Guidelines+1
figma-ds-rules
Generate design system rules that encode codebase conventions for AI agents. Produces CLAUDE.md, AGENTS.md, or .cursor/rules files that guide consistent component generation and token usage.
IMPORTANT: Before creating any component, check (1) get_code_connect_map for existing Figma mappings, (2) src/components/ui/ for shadcn/ui primitives, (3) .memoire/specs/ for existing specs.
CRAFTFIGMA MCP
120
Figma Create File
Create new Figma design or FigJam files via MCP. Resolves plan keys, scaffolds files in drafts, and prepares them for canvas operations with downstream skills.
Plan ResolutionFile ScaffoldingFigJam Boards+1
figma-create-file
Create new Figma design or FigJam files via MCP. Resolves plan keys, scaffolds files in drafts, and prepares them for canvas operations with downstream skills.
Plan ResolutionFile ScaffoldingFigJam BoardsPipeline Handoff
memi notes install figma-create-file
CONTENT PREVIEW
Files are created in the user’s drafts folder for the selected plan. Only "design" and "figJam" editor types are supported. Always load the figma-use skill before making canvas operations on the new file.
CRAFTNEW IN 0.17
35
Motion Performance
Motion performance pack for 60fps UI animation, reduced-motion fallbacks, render budget checks, and video export QA.
60fps BudgetReduced MotionRender QA+1
motion-performance
Motion performance pack for 60fps UI animation, reduced-motion fallbacks, render budget checks, and video export QA.
ACTIVATES ONmotion-videoFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
60fps BudgetReduced MotionRender QAVideo Export
memi notes install motion-performance
CONTENT PREVIEW
Use this Note to keep UI motion and generated video projects fast, accessible, and consistent with the design system. Pair with animation-craft for in-product motion and with remotion-video or hyperframes-video when shipping rendered scenes.
RESEARCH
894
Competitive Intel
Market analysis frameworks, feature matrices, UX benchmarking, strategic positioning, and trend synthesis. Turn competitor data into actionable design decisions.
Market AnalysisFeature MatricesUX Benchmarking+2
competitive-intel
Market analysis frameworks, feature matrices, UX benchmarking, strategic positioning, and trend synthesis. Turn competitor data into actionable design decisions.
Gap Analysis: For each feature where we = "none" and any competitor = "full", Priority = feature_weight * count_of_competitors_with_full. Sort by priority descending. Top 5 = critical gaps to address.
RESEARCH
715
Web Research
Structured web research with topic exploration, content fetching, findings extraction, cross-source validation, and gap analysis for design research workflows.
Structured web research with topic exploration, content fetching, findings extraction, cross-source validation, and gap analysis for design research workflows.
Multi-Angle Queries: Cover the topic from multiple perspectives. 5-7 queries, each targeting a different angle. Example: Form validation research - query best practices, abandonment causes, one-page vs multi-step, guest checkout impact, mobile patterns.
RESEARCH
625
Interview Research
User interview and transcript analysis. Parses speaker-labeled transcripts, extracts pain points and goals, runs sentiment analysis, and generates research-grounded personas.
Transcript ParsingPain Point ExtractionSentiment Analysis+1
interview-research
User interview and transcript analysis. Parses speaker-labeled transcripts, extracts pain points and goals, runs sentiment analysis, and generates research-grounded personas.
ACTIVATES ONresearch-to-dashboardFREEDOMhigh
ALL SKILLS
Transcript ParsingPain Point ExtractionSentiment AnalysisPersona Generation
memi notes install interview-research
CONTENT PREVIEW
When time is limited, extract insights in this order: (1) Pain points - highest signal, most actionable, (2) Goals and motivations - frame the problem space, (3) Current behaviors and workarounds - reveal unmet needs.
RESEARCH
323
Usability Testing
Usability testing frameworks with test planning, task scenarios, moderated and unmoderated protocols, analysis frameworks, and insight synthesis pipelines.
Test PlanningTask ScenariosProtocol Design+2
usability-testing
Usability testing frameworks with test planning, task scenarios, moderated and unmoderated protocols, analysis frameworks, and insight synthesis pipelines.
ACTIVATES ONresearch-to-dashboardFREEDOMhigh
ALL SKILLS
Test PlanningTask ScenariosProtocol DesignAnalysis FrameworksInsight Synthesis
memi notes install usability-testing
CONTENT PREVIEW
Write scenarios, not instructions. Say "Find a book under $30" not "Click the search bar, type books, click filter." Avoid leading language. Do not mention UI elements by name. Include realistic context.
RESEARCH
289
Data Synthesis
Qualitative and quantitative data synthesis with affinity mapping, thematic analysis, statistical frameworks, insight triangulation, and research-to-spec pipelines.
Qualitative and quantitative data synthesis with affinity mapping, thematic analysis, statistical frameworks, insight triangulation, and research-to-spec pipelines.
Phase 1 Open Coding: Read through all data and assign descriptive codes to each meaningful segment. Code everything relevant, skip nothing on first pass. Use gerunds or descriptive phrases. Target 30-80 unique codes from a typical study.
RESEARCH
138
Accessibility Audit
WCAG 2.2 AA/AAA compliance auditing backed by real contrast computation. The v0.5 engine computes actual contrast ratios from hex tokens, validates touch targets, checks focus styles, and produces 0-100 accessibility scores with WCAG level determination.
Color ContrastKeyboard NavigationScreen Reader+2
accessibility-audit
WCAG 2.2 AA/AAA compliance auditing backed by real contrast computation. The v0.5 engine computes actual contrast ratios from hex tokens, validates touch targets, checks focus styles, and produces 0-100 accessibility scores with WCAG level determination.
ACTIVATES ONdesign-reviewFREEDOMread-only
ALL SKILLS
Color ContrastKeyboard NavigationScreen ReaderFocus ManagementSemantic Structure
memi notes install accessibility-audit
CONTENT PREVIEW
WCAG 2.2 contrast: compute relative luminance per sRGB, then (lighter + 0.05) / (darker + 0.05). AA needs 4.5:1 for normal text, 3:1 for large. AAA needs 7:1. Touch targets: 44x44px minimum per criterion 2.5.8.
RESEARCHNEW IN 0.17
37
Browser Research Agent
Run browser-backed research agents that capture evidence, citations, and source freshness - designed for Codex/Hermes/Claude flows.
Web EvidenceCitation CaptureSource Comparison+1
browser-research-agent
Run browser-backed research agents that capture evidence, citations, and source freshness - designed for Codex/Hermes/Claude flows.
ACTIVATES ONbrowser-researchFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
Web EvidenceCitation CaptureSource ComparisonFreshness Checks
memi notes install browser-research-agent
CONTENT PREVIEW
Browser-backed agents must capture (1) URL + retrieval timestamp, (2) the supporting passage verbatim, and (3) at least one comparison source per claim. Use citations as evidence links in research synthesis and Mermaid Jam exports.
CONNECT
1,228
Notion Sync
Notion integration for syncing research insights, specs, and project context between Notion databases and the registry. Handles rich content and database schemas.
Database SyncResearch ImportSpec Export+1
notion-sync
Notion integration for syncing research insights, specs, and project context between Notion databases and the registry. Handles rich content and database schemas.
The sync is idempotent. Every record carries a notion_page_id on the Memoire side and a memoire_spec_id custom property on the Notion side. Conflict resolution follows a last-writer-wins strategy with optional manual review.
CONNECT
1,105
Linear Sync
Linear integration for syncing design issues, component specs, and project tracking. Bi-directional with conflict resolution.
Issue SyncSpec TrackingProject Mapping+1
linear-sync
Linear integration for syncing design issues, component specs, and project tracking. Bi-directional with conflict resolution.
Every design task in Linear maps to a spec (component, page, dataviz, design, or IA). When a spec progresses through its lifecycle, the corresponding Linear issue is updated automatically.
CONNECT
669
Survey Connect
Survey platform integration for importing responses from Typeform, Google Forms, SurveyMonkey, Airtable, and CSV exports into the research pipeline.
Typeform ImportGoogle FormsSurveyMonkey+2
survey-connect
Survey platform integration for importing responses from Typeform, Google Forms, SurveyMonkey, Airtable, and CSV exports into the research pipeline.
Typeform API: Generate a personal access token at admin.typeform.com/account#/section/tokens. Required scopes: forms:read - access form definitions, responses:read - access response data.
CONNECT
630
Self-Improving Agent
Closed-loop learning system with goal tracking, capability evolution, and autonomous improvement cycles. In v0.5, integrates with the multi-agent registry and task queue for distributed self-improvement across agent roles.
Closed-loop learning system with goal tracking, capability evolution, and autonomous improvement cycles. In v0.5, integrates with the multi-agent registry and task queue for distributed self-improvement across agent roles.
Memoire injects these files into every session: CLAUDE.md, notes/, learnings/. When learnings prove broadly applicable, promote them to project files. If a workflow becomes stable, encode it in a built-in note or skill.
CONNECTFIGMA MCP
176
Figma Code Connect
Map Figma design components to codebase implementations using Code Connect. Discovers unmapped components, scans for matching code, and creates batch mappings.
Component DiscoveryCode MatchingBatch Mapping+1
figma-code-connect
Map Figma design components to codebase implementations using Code Connect. Discovers unmapped components, scans for matching code, and creates batch mappings.
When both sides have changes: Token value changed in Figma only → Accept Figma. Token value changed in code only → Push to Figma. Both changed same token → Prompt user - show diff. New component in Figma → Pull + auto-spec.
CONNECTFIGMA MCP
87
Figma Web Capture
Import live webpages or DOM elements into editable Figma layers using the html-to-design bookmarklet.
HTML-to-DesignDOM CaptureEditable Layers+1
figma-web-capture
Import live webpages or DOM elements into editable Figma layers using the html-to-design bookmarklet.
Open the target page in a desktop browser. Wait for fonts, images, and lazy-loaded content to finish rendering. Scroll to the exact section you want to capture. Run the bookmarklet that loads capture.js.
CONNECT
103
Codex Ops
Codex and Claude operating workflow. JSON-first commands, commit hygiene, review loops, and safe autonomous repo operation.
JSON CommandsCommit HygieneReview Loops+1
codex-ops
Codex and Claude operating workflow. JSON-first commands, commit hygiene, review loops, and safe autonomous repo operation.
Prefer machine-readable command output whenever Memoire exposes it. Do not scrape terminal prose if a JSON surface exists. Keep each commit scoped to one behavior or one workflow lane. Read the diff before every commit.
CONNECT
533
Docker Environments
Docker-aware Mémoire operation. Covers Figma bridge port-forwarding, CI/CD headless WCAG and spec pipelines, shared MCP server as a team service, agent workers as containers, and devcontainer setup.
CI/CD PipelineShared MCP ServerAgent Containers+2
docker-environments
Docker-aware Mémoire operation. Covers Figma bridge port-forwarding, CI/CD headless WCAG and spec pipelines, shared MCP server as a team service, agent workers as containers, and devcontainer setup.
The Figma plugin runs on the host machine - not in the container. Map ports 9223-9232 explicitly and bind the bridge to 0.0.0.0. For CI, set MEMOIRE_HEADLESS=true and MEMOIRE_NO_BRIDGE=true. Run memi audit --wcag --exit-code to gate PRs.
CONNECTNEW IN 0.17
37
Agent Memory Profiles
Design durable user, project, persona, and task memory profiles for local agents.
Profile SchemaPersona FilesProject Memory+1
agent-memory-profiles
Design durable user, project, persona, and task memory profiles for local agents.
ACTIVATES ONagent-memory-profileFREEDOMhighSOURCENEW IN 0.17
A messaging gateway turns agent output into channel-aware deliveries. Encode the routing contract once (channel, audience, redaction policy, retry window) so any agent kit can publish without re-implementing platform plumbing.
CONNECTNEW IN 0.17
37
Agent Session Checkpoints
Capture restartable agent sessions with checkpoints, rollback notes, and resume metadata.
CheckpointingRollbackResume Metadata+1
agent-session-checkpoints
Capture restartable agent sessions with checkpoints, rollback notes, and resume metadata.
ACTIVATES ONagent-checkpointFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
CheckpointingRollbackResume MetadataSession History
memi notes install agent-session-checkpoints
CONTENT PREVIEW
A checkpoint is a labeled snapshot of agent state plus the prompt that produced it. Resume from a checkpoint by replaying the prompt against the same session store - never by hand-editing the JSONL log.
CONNECTNEW IN 0.17
37
Agent Skill Migration
Move skills, memories, context files, and agent rules between Memoire and peer agent ecosystems (Codex, Hermes, OpenClaw).
Skill ImportMemory ExportCodex/Hermes/OpenClaw+1
agent-skill-migration
Move skills, memories, context files, and agent rules between Memoire and peer agent ecosystems (Codex, Hermes, OpenClaw).
ACTIVATES ONagent-skill-migrationFREEDOMhighSOURCENEW IN 0.17
When porting a skill: keep the original system prompt verbatim, normalize tool names to the target ecosystem, then add a Memoire `activateOn` signal so router heuristics still work after the move.
CONNECTNEW IN 0.17
37
Apple Desktop Automation
Prepare macOS agents to work with Apple apps, local files, permissions, and desktop automation (Notes, Reminders, iMessage).
Prepare macOS agents to work with Apple apps, local files, permissions, and desktop automation (Notes, Reminders, iMessage).
ACTIVATES ONmacOS-automationFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
macOS PermissionsAppleScriptApple Notes/RemindersDesktop Control
memi notes install apple-desktop-automation
CONTENT PREVIEW
macOS agents must request Automation, Accessibility, and Full Disk Access privileges explicitly per app. Bundle the prompts behind a single `memi setup macos` flow so users grant access once instead of mid-session.
CONNECTNEW IN 0.17
37
Approval & Sandbox Policies
Define approval, sandbox, filesystem, and command policy for local autonomous agents.
Approval FlowsSandbox ModesCommand Allowlist+1
approval-sandbox-policies
Define approval, sandbox, filesystem, and command policy for local autonomous agents.
ACTIVATES ONagent-approvalFREEDOMhighSOURCENEW IN 0.17
Default to a guarded sandbox: read-only filesystem, allowlisted commands, network off. Promote to write or shell only when an explicit user approval landed in the same turn - never inherit elevated permissions across sessions.
CONNECTNEW IN 0.17
37
Cron Agent Workflows
Turn repeatable agent work into scheduled jobs with explicit outputs and safe permissions.
Every scheduled run should produce a single JSONL artifact and exit cleanly. Treat retry as the user’s job - agents that retry themselves on a cron silently double their cost and surface area.
CONNECTNEW IN 0.17
37
Gateway Ops & Observability
Observe and debug local agent gateways, daemons, health probes, logs, and usage costs.
Health ProbesDaemon StatusCost Tracking+1
gateway-ops-observability
Observe and debug local agent gateways, daemons, health probes, logs, and usage costs.
ACTIVATES ONgateway-opsFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
Health ProbesDaemon StatusCost TrackingLog Analysis
memi notes install gateway-ops-observability
CONTENT PREVIEW
A healthy gateway exposes three signals: liveness (process up), readiness (downstream models reachable), and quota (cost-window remaining). Treat any of the three flipping as a scheduling event, not a paging event.
CONNECTNEW IN 0.17
37
Hermes Agent Bridge
Run Hermes Agent from Memoire with CLI, skills, toolsets, worktree, and session recipes.
Hermes CLISkill PreloadWorktree Sessions+1
hermes-agent-bridge
Run Hermes Agent from Memoire with CLI, skills, toolsets, worktree, and session recipes.
ACTIVATES ONhermes-bridgeFREEDOMhighSOURCENEW IN 0.17
Memoire installs `memoire-design-tooling` under ~/.hermes/skills. Activate it inside a Hermes session by referencing the skill name explicitly - Hermes does not auto-load skills the way Codex does.
CONNECTNEW IN 0.17
37
MCP Server Studio
Expose Memoire Studio context, tools, and artifacts to other agents through MCP.
MCP WiringTool ExposureArtifact Sharing+1
mcp-server-studio
Expose Memoire Studio context, tools, and artifacts to other agents through MCP.
ACTIVATES ONmcp-studioFREEDOMhighSOURCENEW IN 0.17
`memi mcp start --no-figma` is the registry-safe entrypoint. External agents can list 18 Memoire tools and 3 resources without requiring a live Figma bridge - useful for crawlers and headless inspections.
CONNECTNEW IN 0.17
37
Model Router Diagnostics
Diagnose provider auth, model routing, quota windows, and fallback behavior before agent runs.
Provider AuthQuota WindowsFallback Models+1
model-router-diagnostics
Diagnose provider auth, model routing, quota windows, and fallback behavior before agent runs.
ACTIVATES ONmodel-routingFREEDOMhighSOURCENEW IN 0.17
Run `memi simulate models --json` before any swarm cohort. The matrix shows which provider is reachable, which fallback model fires, and what reasoning effort is currently configured - preventing silent quality drops mid-run.
CONNECTNEW IN 0.17
37
Multi-Agent Kanban
Coordinate multiple local agents through a durable board with owner, state, and evidence fields.
Board SchemaOwner RoutingEvidence Fields+1
multi-agent-kanban
Coordinate multiple local agents through a durable board with owner, state, and evidence fields.
ACTIVATES ONmulti-agent-boardFREEDOMhighSOURCENEW IN 0.17
Treat the kanban as the single source of truth for inter-agent coordination. Each card has owner, state, and evidence link - agents that change state without writing evidence get rejected on the next sync.
CONNECTNEW IN 0.17
37
OpenClaw Agent Bridge
Connect Memoire to OpenClaw onboarding, agents, gateway health, MCP, and approval flows.
OpenClaw SetupGateway RoutingOpenClaw MCP+1
openclaw-agent-bridge
Connect Memoire to OpenClaw onboarding, agents, gateway health, MCP, and approval flows.
ACTIVATES ONopenclaw-bridgeFREEDOMhighSOURCENEW IN 0.17
OpenClaw expects skills under `<workspace>/skills/`. Memoire installs `memoire-design-tooling` there and registers itself in the gateway so OpenClaw approvals route through Memoire’s permission mode.
CONNECTNEW IN 0.17
37
Secure Secrets for Agents
Handle agent credentials with env refs, redaction, least privilege, and non-leaking prompts.
Env RefsRedactionLeast Privilege+1
secure-secrets-for-agents
Handle agent credentials with env refs, redaction, least privilege, and non-leaking prompts.
ACTIVATES ONsecret-handlingFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
Env RefsRedactionLeast PrivilegePrompt Hygiene
memi notes install secure-secrets-for-agents
CONTENT PREVIEW
Never paste an API key into a prompt. Reference env vars by name; the Memoire daemon resolves them at call time and redacts them in transcripts. Treat any agent that requests a literal secret as a security event.
GENERATE
1,676
Flutter Generator
Flutter and Dart widget generation from specs. Produces Material 3 widgets with Riverpod state management, responsive layouts, and platform-adaptive patterns.
Material 3 WidgetsRiverpod StateResponsive Layouts+1
flutter-gen
Flutter and Dart widget generation from specs. Produces Material 3 widgets with Riverpod state management, responsive layouts, and platform-adaptive patterns.
ACTIVATES ONcomponent-creationFREEDOMhigh
ALL SKILLS
Material 3 WidgetsRiverpod StateResponsive LayoutsPlatform Patterns
memi notes install flutter-gen
CONTENT PREVIEW
Generates Material 3 themed widgets with Riverpod state management, responsive layouts, custom painting for dataviz, GoRouter navigation, and full test coverage. Follows Effective Dart and atomic design principles.
GENERATE
383
Vue Generator
Vue 3 Composition API code generation. Transforms component specs into Vue SFCs with script setup, Tailwind styling, and composable patterns.
Vue 3 SFCsComposition APITailwind Styling+1
vue-gen
Vue 3 Composition API code generation. Transforms component specs into Vue SFCs with script setup, Tailwind styling, and composable patterns.
Atoms are self-contained. No imports from other components. Props in, events out. Molecules compose 2-5 atom components. Keep logic minimal - presentation layer only. Organisms use composables for complex state.
GENERATEFIGMA MCP
183
Figma Implement Design
Structured workflow for translating Figma designs into production-ready code with 1:1 visual parity.
Design ExtractionToken MappingCode Generation+1
figma-implement-design
Structured workflow for translating Figma designs into production-ready code with 1:1 visual parity.
Tailwind to NativeWind: Web Tailwind classes map 1:1 in NativeWind 4.x. Exceptions: hover:, focus-visible: - replace with active: states. Grid: Use flex layout - RN has no CSS Grid. gap-* works in NativeWind, prefer over margin hacks.
GENERATENEW IN 0.17
43
Hyperframes Video
Hyperframes pack for HTML-first motion videos with lintable frames, previews, and render command prep.
HTML ScenesFrame LintingPreview Server+1
hyperframes-video
Hyperframes pack for HTML-first motion videos with lintable frames, previews, and render command prep.
ACTIVATES ONmotion-videoFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
HTML ScenesFrame LintingPreview ServerRender Handoff
memi notes install hyperframes-video
CONTENT PREVIEW
Use Hyperframes when the video can be expressed as a deterministic HTML scene - a layered DOM tree with timed transitions. The Note ships render commands for the local preview server and the production frame export.
GENERATENEW IN 0.17
43
Remotion Video
Remotion video production pack for React-based product videos, previews, and renderer handoff.
React VideosComposition SetupPreview Studio+1
remotion-video
Remotion video production pack for React-based product videos, previews, and renderer handoff.
ACTIVATES ONmotion-videoFREEDOMhighSOURCENEW IN 0.17
Use Remotion when the output should be a React-coded video project. Keep one Composition per scene, defer typography and easing to design tokens, and hand off the rendered MP4 through the Memoire downloads store.
GENERATENEW IN 0.17
34
Website to Video
Website-to-video capture pack for turning product pages, flows, and browser states into structured video scenes.
Page CaptureFlow RecordingScene Structuring+1
website-to-video
Website-to-video capture pack for turning product pages, flows, and browser states into structured video scenes.
ACTIVATES ONmotion-videoFREEDOMhighSOURCENEW IN 0.17
ALL SKILLS
Page CaptureFlow RecordingScene StructuringBrowser State
memi notes install website-to-video
CONTENT PREVIEW
Use this Note when Studio needs to turn a website, app route, or browser flow into a structured video scene - captures viewport, scroll, and interaction states without re-implementing each page in Remotion or Hyperframes.
HOW IT WORKS
01
Install
Run memi notes install <name> to download a Note into your project's .memoire/notes/ directory.
02
Auto-activate
Notes are loaded at engine init and injected into agent prompts when their activateOn context matches the classified intent.
03
Deep knowledge
Each Note contains hundreds of lines of real expertise, not templates. They transform Mémoire into a domain specialist.
04
Create your own
Run memi notes create <name> to scaffold a new Note. Add your domain knowledge in markdown, set the activation context, and it's ready.
CLI REFERENCE
memi notes install <source>Install a note from local path or github:user/repo
memi notes listShow all installed notes with status and category
memi notes remove <name>Uninstall a note from your project
memi notes create <name>Scaffold a new note with manifest and skill template
memi notes info <name>Show detailed note information and skills