Relay equips your AI coding assistant with a multidisciplinary software & brand engineering system — turning it into a Senior QA Engineer, Security Auditor, Brand Strategist, Tech Lead, or Release Manager on demand.

Relay provides a suite of highly optimized Agent-Native Skills & Workflows that install directly into your AI developer assistant (Claude Code, Google Antigravity, Cursor, or GitHub Copilot) as native slash commands.

From scientific brand strategy & advertising psychology to OWASP security audits & release gating, Relay provides structured decision frameworks for the entire software creation lifecycle.

🚀 Quick Start

Run the automated installer inside any project directory:

npx @akshatdev/relay install-skills

By default, Relay automatically installs all skills locally (for the current workspace) and globally (across your machine) for all detected assistant providers.

⚡ Assistant Integrations

| Assistant | Local Slash Command / Workflow Path | Global Installation Path | | :--- | :--- | :--- | | Claude Code | /relay-<skill> via .claude/commands/ | ~/.claude/commands/ | | Google Antigravity | /relay-<skill> via .agents/workflows/ | ~/.gemini/config/global_workflows/ | | Cursor | Persisted rules under .cursor/rules/ | Persistent instructions | | GitHub Copilot | Appended to .github/copilot-instructions.md | Persistent prompts |

🎨 Brand & Product Experience System

💎 /relay-brand (/relay-brand-experience, /relay-ibe) — Integrated Brand Experience

A 22-stage design operating system that transforms a business idea or product into a cohesive digital experience by fusing Marketing Strategy, Human Psychology, Visual/Brand Identity, and Web/Interaction Design.

Grounded in peer-reviewed cognitive neuroscience, advertising psychology, and visual storytelling research:

• Norman’s Tri-Level Emotional Processing (Visceral, Behavioral, Reflective)
• Ramachandran’s Peak Shift Principle & Neuroaesthetics
• Fauconnier & Turner’s Conceptual Blending for Dual-Storytelling Logos
• Jonah Berger’s STEPPS & Arousal Theory for Virality
• Mandler’s Schema Incongruity & Processing Fluency

BRAND → HUMAN INSIGHT → MARKETING STRATEGY → PSYCHOLOGY → CREATIVE CONCEPT → VISUAL IDENTITY → LOGO / SYMBOL → WEB EXPERIENCE → INTERACTION → EMOTION → MEMORY → BEHAVIOR → CONVERSION → BUSINESS OUTCOME

🛠️ Code Audit & Quality Skills

Every skill is structured as an agent-native instruction set and can be invoked natively using slash commands in your AI assistant's chat interface.

📋 /relay-qa — Full Project QA Audit

Evaluates code quality, security, performance, accessibility, and release readiness to calculate a weighted score [0-100] and output a definitive SHIP / WARN / HOLD decision.

| Category | Weight | Focus Areas | | :--- | :---: | :--- | | Security | 30% | OWASP Top 10, secrets in source code, input validation | | Testing | 20% | Coverage, assertion quality, focused/skipped tests | | Architecture | 20% | Layering, cohesion, circular dependencies, God files | | Performance | 10% | Bundle tree-shaking, lazy-loading, N+1 query patterns | | Accessibility | 10% | Semantic HTML, ARIA labels, keyboard focus states | | Release Readiness | 10% | Version verification, changelogs, configuration templates |

🛡️ /relay-security — OWASP Security Audit

An intensive security assessment mapped directly against OWASP Top 10 vulnerabilities (including injection, broken access control, cryptographic failures, and insecure design). Returns SECURE / REVIEW NEEDED / VULNERABLE.

🚀 /relay-release — Release Readiness Check

Gates your code before shipping. Verifies dependency lockfile health, floating production dependencies, changelog entries, version increments, container user privileges, and environment variable templates. Returns a clear GO / NO-GO verdict.

⚡ /relay-performance — Performance Audit

Identifies frontend and backend bottlenecks. Audits asset lazy-loading, dynamic importing, monolithic library tree-shaking, HTTP caching headers, and database query pagination.

🧪 /relay-testing — Test Quality Verification

Audits test suite reliability and coverage. Flags accidentally committed .only/.skip focus markers, empty test files, dummy placeholder assertions, and brittle over-mocking patterns.

🏗️ /relay-architecture — Code Architecture Audit

Evaluates structural health and separation of concerns. Flags monolithic dependencies, circular imports, deep relative paths (../../../../), and bloated files exceeding 1000 lines of code.

💻 CLI Usage & CI Integration

Relay maintains full CLI compatibility for automated CI/CD pipelines and headless workflows:

# Print current Relay CLI version
relay version

# Run a structured review and export JSON results
relay review --format json --cwd .

🤝 Contributing

Contributions are welcome! Check out our Contributing Guide to get started.

MIT License © Relay Contributors