Agent ecosystem · 25 specifications
Agents are infrastructure.
Designers remain accountable.
Every agent has defined inputs, outputs, model needs, example prompts, skills trained, and a visible line it must never cross.
Requirements Agent
Runs structured, asynchronous stakeholder intake and synthesizes it into a draft charter
Research Agent
Ingests and clusters unstructured feedback and research at volume
Persona Agent
Drafts evidence-based personas from clustered research for human validation
JTBD Agent
Constructs Jobs-to-be-Done statements from research and persona data
Competitor/Market Agent
Maps the competitive landscape structurally
Insight Synthesis Agent
Drafts candidate problem statements and HMW framings for human selection
Journey Agent
Maps current-state and future-state user journeys from research and persona data
Ideation Agent
Generates high-volume structural and conceptual variation against the problem statement
Design Critique Agent
Evaluates concepts against brand, feasibility signals, and ethical constraints
IA Agent
Generates initial structure, taxonomy, and cognitive-load diagnostics
UX Writer Agent
Drafts microcopy and navigation labels
Localization Agent
Adapts taxonomy, labeling, and copy for target-market language and cultural context
Interaction/Flow Agent
Auto-links screens and generates standard micro-interactions
Design System Agent
Applies design-system tokens/components at scale and monitors drift
Figma Agent
Bridges design files and code bidirectionally; shared with Stage 08
Motion Agent
Generates standard motion/easing patterns, distinct from bespoke choreography
Multimodal Interaction Agent
Designs and evaluates voice, and other non-visual interaction modalities
UX QA Agent
Simulates usability heuristics and calculates efficiency metrics
Accessibility Agent
Runs automated WCAG audits continuously
Prototype Agent
Runs AI-moderated and synthetic pre-interviews at scale
Translation Agent
Generates semantic code and tagged specs from validated designs
Security/Privacy Scan Agent
Pre-scans AI-generated code for common security/privacy issues, ahead of mandatory human review
Analytics Agent
Monitors live telemetry continuously and flags anomalies
Prompt Optimizer
Improves prompt patterns across all other agents based on observed output quality
Governance Agent
Maintains the audit trail and compliance thread across every stage