emote-behavioral-contracts

April 1, 2026 · View on GitHub

A Figma MCP skill that reads Emote behavioral annotation components from a Figma frame and produces implementation-ready behavioral contracts for AI-driven interfaces.

Design systems tell AI how to build. Emote tells AI how to behave.

What this skill does

When a designer annotates an AI-driven interface using the Emote Design System annotation kit, this skill reads those annotations via the Figma MCP server and outputs:

A JSON component manifest for use in component implementations
A system prompt block for use in AI agent specifications
A gap report if required patterns are absent from the annotated frame

The skill bridges the Emote annotation layer — written by designers — and the implementation layer consumed by engineers and AI agents.

Emote is a behavioral specification framework for AI systems. It defines six trust moments — recurring points in an interaction where AI behavior carries explicit obligations — and a pattern language and token vocabulary that make those obligations specifiable, reviewable, and implementable.

The six patterns are:

ID	Pattern	When it applies
P01	Expectation Setting	Before the system acts
P02	Ambiguity Detection	When intent is unclear
P03	Interpretive Support	When the system must restore shared understanding
P04	Consent Confirmation	Before a boundary-crossing or irreversible action
P05	Repair and Apology	When the system causes harm or confusion
P06	State Reorientation	After a rupture or state change

Full pattern contracts, token definitions, and implementation guidance at emote.dev.

Requirements

Figma MCP server connected to your MCP client
A Figma frame annotated with Emote components from the Emote Design System annotation kit
Claude Code, Cursor, or Codex

Installation

Claude Code

.claude/skills/emote-behavioral-contracts/SKILL.md

Place the emote-behavioral-contracts/ folder at the above path in your repo. In Claude Code, invoke with:

/emote-behavioral-contracts

Cursor

In Cursor chat:

/create-skill

Then follow the prompts, or manually place the folder in your workspace skills location.

Codex

$skill-creator

Follow the prompts to register the skill.

Usage

Open or select an annotated Figma frame in your MCP client
Invoke the skill: /emote-behavioral-contracts
The skill will:
- Detect Emote annotation components in the frame
- Inventory active patterns and tokens
- Identify the safe failure mode
- Output a JSON component manifest and system prompt block
- Report any behavioral gaps if GapMarkers are present

Example output

For a frame annotated with P01 (Expectation Setting) and P02 (Ambiguity Detection):

{
  "emote": {
    "component": "agent-response-bar",
    "patterns": [
      {
        "pattern": "P01_expectation_setting",
        "tokens": [
          "behavior.set_expectations_early",
          "behavior.state_time_and_steps",
          "behavior.clarify_agency_boundaries"
        ]
      },
      {
        "pattern": "P02_ambiguity_detection",
        "trigger": "multi_interpretation_detected",
        "tokens": [
          "behavior.pause_when_uncertain",
          "behavior.clarify_before_action",
          "behavior.name_risk_transparently",
          "behavior.delay_irreversible_actions"
        ],
        "safe_failure": "stop_state_known_route_human"
      }
    ]
  }
}

Annotation components the skill reads

The skill recognizes these components from the Emote Design System annotation kit:

Component	What it means
`FrameBadge`	Frame is under behavioral spec
`PatternStamp`	Active trust moment for a flow or screen
`PatternStampTag`	Compact pattern identifier on a single element
`BehaviorCallout`	Trust moment, trigger condition, or safe failure mode
`TokenTag`	Single behavioral token obligation on an element
`TrustMomentCallout`	Full behavioral contract: pattern, promise, token set
`GapMarker`	Region where a required pattern is absent
`GapMarkerTag`	Numbered reference to a GapMarker

Emote — behavioral specification framework
Emote Design System — Figma community file with annotation kit, pattern reference, and AI handoff examples
Emote Doctrine — the principles behind the framework
Emote Patterns — full pattern contracts
Emote Tokens — canonical behavioral token vocabulary

License

MIT

Created by Roger Branon Rodriguez