⚡ StringRay AI

Enterprise AI Orchestration Framework for OpenCode/Claude Code

Intelligent Multi-Agent Coordination with 99.6% Systematic Error Prevention

StringRay extends OpenCode/Claude Code with intelligent multi-agent orchestration, Codex compliance validation, and enterprise-grade security. It automatically routes tasks to specialized agents based on complexity and provides systematic error prevention.

What is StringRay?

StringRay is a one-command level-up for OpenCode. Instead of installing OpenCode first, then adding StringRay, just run:

npx strray-ai install

This single command:

Detects if OpenCode is installed
Auto-installs OpenCode if missing
Layers on the full StringRay kernel (Codex, orchestrator, enforcer, processors, MCP, reflections)
Installs 44 framework skills
Sets up the skills registry with 10 curated community sources
Adds CLI commands for agent publishing, skills management, and status

Goal: Any developer can run one command and instantly get a production-grade, governed agent runtime.

Who is it for?

Developers using OpenCode or Claude Code who want AI-assisted development
Teams needing consistent code quality and error prevention
Enterprises requiring security, compliance, and audit capabilities

🚀 Quick Start

# Install StringRay (auto-configures OpenCode on install)
npm install strray-ai

# That's it! StringRay is now active.
# Restart OpenCode/Claude Code to load the plugin.

What happens during install?

Copies OpenCode configuration files to your project
Configures 25 agents with proper capabilities
Sets up Codex enforcement rules
Enables webhook triggers for CI/CD integration
Ready to use with Claude Code immediately

Standalone Mode (No OpenCode Required)

For use with Hermes Agent without OpenCode:

npx strray-ai install --standalone

This installs only the MCP servers - no OpenCode dependency.

✨ Features

🤖 25 Specialized Agents - Autonomous agents that read/write code, run commands, and enforce compliance
📏 99.6% Error Prevention - Universal Development Codex (60 terms)
⚡ 44 Framework Skills + 10 curated community sources (170+ additional skills available)
🛡️ Enterprise Security - Comprehensive validation and scanning
📦 Skills Registry - Browse and install community skills from GitHub repos
🔄 Complexity-Based Routing - Intelligent task delegation
🔌 Webhook Integration - GitHub, GitLab, Bitbucket, Stripe
✅ 2311 Tests - Production-ready with comprehensive test coverage
🧩 Standalone MCP Servers - Works with Hermes Agent without OpenCode

🤖 Available Agents

Agent	Purpose
`@enforcer`	Codex compliance & error prevention
`@orchestrator`	Complex multi-step task coordination
`@architect`	System design & technical decisions
`@security-auditor`	Vulnerability detection
`@code-reviewer`	Quality assessment
`@refactorer`	Technical debt elimination
`@testing-lead`	Testing strategy & coverage
`@bug-triage-specialist`	Error investigation
`@researcher`	Codebase exploration

Note: StringRay auto-configures all agents during installation. To customize agent settings, see the Agent Configuration Guide.

View all 25 agents →

📦 OpenClaw Integration

StringRay integrates with OpenClaw - a self-hosted AI gateway that connects messaging platforms (WhatsApp, Telegram, Discord, Slack) to AI coding agents.

What It Does

WebSocket Connection: Connect to OpenClaw Gateway at ws://127.0.0.1:18789
Skill Invocation: OpenClaw skills invoke StringRay agents via HTTP API (port 18431)
Tool Events: Forward tool.before/tool.after events to OpenClaw for real-time tracking
Offline Buffering: Events queued when disconnected, sent on reconnect

Quick Setup

# Configure in .opencode/openclaw/config.json
{
  "gatewayUrl": "ws://127.0.0.1:18789",
  "authToken": "your-device-token",
  "deviceId": "your-device-id",
  "apiServer": { "enabled": true, "port": 18431 },
  "hooks": { "enabled": true, "toolBefore": true, "toolAfter": true }
}

# Initialize in code
import { initializeOpenClawIntegration } from 'strray-ai';
const integration = await initializeOpenClawIntegration();

See OpenClaw Integration Guide for details.

🔮 Hermes Agent Integration

StringRay's MCP servers work as native tools in Hermes Agent — a standalone AI coding agent with its own runtime, not dependent on OpenCode or Claude Code. This gives you the full StringRay toolset (code analysis, linting, security scanning, orchestration, state management) inside any Hermes session.

What You Get

Hermes discovers 10 StringRay MCP servers automatically. Each server exposes tools that Hermes can call directly — no prompts, no proxies, no agent delegation overhead.

MCP Server	Tools Exposed	What It Does
`strray-architect-tools`	codebase_structure, dependency_analysis, context_analysis, architecture_assessment	Analyze project structure, dependencies, patterns, and architectural health
`strray-auto-format`	auto_format, format_check	Prettier + ESLint + TypeScript formatting and validation
`strray-enforcer`	rule_validation, codex_enforcement, quality_gate_check, run_pre_commit_validation	Codex compliance, quality gates, pre-commit validation
`strray-estimation`	validate_estimate, start_tracking, complete_tracking, get_accuracy_report	Task estimation with calibration and accuracy tracking
`strray-framework-help`	strray_get_capabilities, strray_get_commands, strray_explain_capability	Framework reference, agent docs, capability lookups
`strray-lint`	lint, lint_check	ESLint validation with auto-fix and rule-specific checks
`strray-orchestrator`	orchestrate_task, analyze_complexity, get_orchestration_status, optimize_orchestration, cancel_orchestration	Multi-agent task planning, complexity scoring, parallel optimization
`strray-researcher`	search_codebase, find_implementation, get_documentation	Codebase search, pattern finding, documentation lookup
`strray-security-scan`	security_scan, dependency_audit	Vulnerability scanning and dependency audit
`strray-state-manager`	get_state, set_state, delete_state, list_state, backup_state, restore_state, validate_state	Persistent key-value state with backup/restore

Setup

Install StringRay in your project:

npm install strray-ai

Add the MCP server entries to your Hermes config (~/.hermes/config.yaml):

mcp_servers:
  strray-architect-tools:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/architect-tools.server.js
    timeout: 30
  strray-auto-format:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/auto-format.server.js
    timeout: 30
  strray-enforcer:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/enforcer-tools.server.js
    timeout: 30
  strray-estimation:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/estimation.server.js
    timeout: 30
  strray-framework-help:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/framework-help.server.js
    timeout: 30
  strray-lint:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/lint.server.js
    timeout: 30
  strray-orchestrator:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/orchestrator/server.js
    timeout: 60
  strray-researcher:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/researcher.server.js
    timeout: 60
  strray-security-scan:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/security-scan.server.js
    timeout: 30
  strray-state-manager:
    command: node
    args:
      - ./node_modules/strray-ai/dist/mcps/state-manager.server.js
    timeout: 30

Restart Hermes. The tools will appear with the mcp_strray_ prefix.

Usage Examples

Once connected, Hermes can use the tools directly in conversation:

"Analyze this project's architecture and give me a health score"
"Run a security scan on the codebase"
"Lint all TypeScript files and auto-fix what you can"
"Plan a refactoring of the auth module — break it into steps"
"Track my time: I estimate this feature will take 2 hours"
"Search the codebase for all database query patterns"

How It Differs from OpenCode

	OpenCode Plugin	Hermes MCP
Runtime	OpenCode or Claude Code	Any Hermes session (CLI, Telegram, Discord)
Discovery	Plugin injection via `opencode.json`	MCP protocol via `config.yaml`
Tool Access	Agent-to-agent delegation	Direct tool calls from Hermes
Context	Shared OpenCode session	Full Hermes session with memory
Platforms	Terminal only	CLI, Telegram, Discord, WhatsApp, Slack

Tips

Use absolute paths in args if Hermes runs from a different working directory than your project
Increase timeout for orchestrator and researcher (60s) — they do heavier analysis
The state-manager persists to .strray/state/mcp-state.json — survives Hermes restarts
Auto-format and lint need Prettier/ESLint installed in your project to do real work (otherwise they report what's missing)

📖 Documentation

Guide	Description
Agent Configuration	Copy-paste opencode.json agent setup
Configuration Reference	Complete features.json settings
Agent Documentation	Detailed agent specifications
Universal Codex	60-term codex reference
Troubleshooting	Common issues & solutions

🔧 CLI Tools

StringRay provides CLI utilities for managing and monitoring your installation:

# Core commands
npx strray-ai status              # Check configuration and plugin status
npx strray-ai validate            # Validate installation and dependencies
npx strray-ai capabilities       # Show all available features
npx strray-ai health             # Run health check on framework components
npx strray-ai report             # Generate usage and performance reports

# Agent management
npx strray-ai publish-agent --agent orchestrator  # Package agent for AgentStore

# Skills management
npx strray-ai skill:install              # Show starter packs + available sources
npx strray-ai skill:install agency-agents  # Install from registry (auto-detects format)
npx strray-ai skill:install <github-url>  # Install from any repo
npx strray-ai skill:registry list      # Show all registry sources
npx strray-ai antigravity status       # Show installed skills with licenses

Note: Installation is automatic via npm install strray-ai. The postinstall hook configures everything automatically.

⚙️ Configuration

Default Configuration

StringRay works out of the box with sensible defaults. The npm postinstall hook automatically sets up:

.opencode/
├── agents/         # 24+ agent configurations
├── skills/         # Framework skills
├── strray/
│   ├── codex.json      # Codex rules
│   ├── features.json   # Feature flags
│   └── config.json    # Token/memory management
└── hooks/          # Pre/post processing hooks

Customizing Agents

Edit .opencode/agents/ to customize agent behavior:

# Example: Customize enforcer agent
name: enforcer
maxComplexity: 40  # Only handle simple tasks
temperature: 0.2   # More precise responses
enabled: true

Feature Flags

Edit .opencode/strray/features.json to enable/disable features:

{
  "codexEnforcement": true,
  "agentGovernance": true,
  "analytics": true,
  "webhooks": true
}

Token Management

Edit .opencode/strray/config.json to adjust token limits:

{
  "token_management": {
    "maxPromptTokens": 20000,
    "warningThreshold": 15000
  }
}

See Configuration Reference for full options.

Version Pinning

StringRay supports pinning versions for reproducible installations:

{
  "version_pinning": {
    "strray_ai": "^1.15.0",
    "opencode": "^2.14.0",
    "skills": {
      "antigravity": "latest",
      "impeccable": "latest",
      "openviking": "latest",
      "claude_seo": "latest"
    }
  }
}

Add to .opencode/strray/features.json to pin specific versions.

📁 Project Structure

stringray/
├── src/
│   ├── __tests__/              # Test suites (unit, integration, performance)
│   ├── agents/                 # Agent implementations
│   ├── analytics/              # Pattern analysis & learning
│   ├── cli/                    # CLI commands
│   ├── circuit-breaker/        # Resilience patterns
│   ├── core/                   # Core framework
│   ├── delegation/             # Task routing & delegation
│   ├── enforcement/            # Codex enforcement
│   ├── infrastructure/         # IaC validation
│   ├── integrations/           # External integrations
│   │   ├── base/              # BaseIntegration framework
│   │   └── openclaw/          # OpenClaw integration
│   ├── mcps/                  # MCP server implementations
│   ├── monitoring/            # System monitoring
│   ├── orchestrator/          # Multi-agent orchestration
│   ├── performance/           # Performance optimization
│   ├── plugins/              # Plugin system
│   ├── postprocessor/         # Post-processing pipeline
│   ├── reporting/             # Report generation
│   ├── security/              # Security systems
│   ├── session/               # Session management
│   ├── test-utils/            # Test utilities and helpers
│   ├── validation/            # Agent config & estimation validators
│   └── jobs/                  # Background job management
├── .opencode/                 # OpenCode configuration
│   ├── agents/               # Agent configs (25 agents)
│   ├── strray/               # StringRay config
│   │   ├── codex.json        # 60-term development codex
│   │   ├── features.json     # Feature flags
│   │   └── config.json       # Token management
│   └── hooks/                # Git hooks
├── skills/                    # StringRay skills
├── docs/                      # Documentation
│   ├── reflections/          # Deep technical reflections
│   └── research/             # Research documents
└── scripts/                   # Build & utility scripts

💬 Usage

# Code quality enforcement
@enforcer analyze this code for issues

# Complex task orchestration  
@orchestrator implement user authentication system

# System design
@architect design database schema for e-commerce

# Security audit
@security-auditor scan for vulnerabilities

🔌 Framework Integration

StringRay integrates with your existing infrastructure via webhooks and APIs:

# CLI tool for integration
npx strray-integration --help

// Programmatic integration
import { StringRayIntegration } from 'strray-ai/integration';

const postProcessor = new PostProcessor(stateManager);
const integration = new StringRayIntegration(postProcessor);

// Express
app.use('/webhooks', integration.getWebhookApp());
app.use('/api/post-process', integration.getAPIApp());

// Fastify
fastify.register(integration.getWebhookRouter(), { prefix: '/webhooks' });
fastify.register(integration.getAPIRouter(), { prefix: '/api/post-process' });

Supported Webhooks:

GitHub (push, PR, issues)
GitLab (push, merge requests)
Bitbucket (push, pull requests)
Stripe (subscriptions, payments)

🎯 Skills

StringRay ships with 44 framework skills and provides a registry of 10 curated community sources with 170+ additional skills.

Skills Registry

Browse and install skills from verified GitHub repositories:

# Show starter packs and available sources
npx strray-ai skill:install

# Install a specific source
npx strray-ai skill:install agency-agents
npx strray-ai skill:install superpowers
npx strray-ai skill:install anthropic-skills

# Install from any GitHub repo (auto-detects format)
npx strray-ai skill:install https://github.com/user/skills-repo

# Manage the registry
npx strray-ai skill:registry list              # Show all sources
npx strray-ai skill:registry add --name X --url Y --desc "..." --license MIT
npx strray-ai skill:registry remove --name X

Starter Packs

Pack	Sources	Skills	Best For
Minimal Viable Power	superpowers, anthropic-skills	20+	Solo devs, quick setup
Full Pro Setup	+ agency-agents, impeccable, minimax	200+	Professional development
Agency/Team Mode	+ gemini-skills, ai-web3-security	220+	Teams, security audits
Specialized	+ vuejs-nuxt, ui-ux-pro-max	230+	Nuxt/Vue, UI/UX work

Registry Sources

Source	Skills	License	Description
agency-agents	170+	MIT	AI agency agent definitions
superpowers	14	MIT	TDD, debugging, code review workflows
anthropic-skills	10+	MIT	Official Anthropic Claude Code skills
antigravity	1300+	MIT	Curated community skills
impeccable	1	Apache 2.0	AI frontend design language
minimax	20+	MIT	Frontend, mobile, shader skills
gemini-skills	10+	Apache 2.0	Official Google Gemini skills
ai-web3-security	10+	MIT	Web3 security auditing
vuejs-nuxt	5+	MIT	Vue.js 3, Nuxt 4+ skills
ui-ux-pro-max	1	MIT	Professional UI/UX design

Impeccable - AI Frontend Design

Impeccable is a design language skill that teaches AI coding assistants professional frontend design:

/audit           # Find issues
/critique       # UX design review
/polish          # Pre-ship refinement
/typeset         # Fix typography
/arrange         # Fix layout & spacing

Framework Skills (30 Built-in)

StringRay includes 30 core skills for orchestration, compliance, architecture, and more — installed to .opencode/skills/ automatically.

License Information

All community skill sources are properly licensed. License files are in licenses/skills/:

Source	License	File
agency-agents	MIT	`licenses/skills/LICENSE.agency-agents`
superpowers	MIT	`licenses/skills/LICENSE.superpowers`
anthropic-skills	MIT	`licenses/skills/LICENSE.anthropic-skills`
antigravity	MIT	`licenses/skills/LICENSE.antigravity`
impeccable	Apache 2.0	`licenses/skills/LICENSE.impeccable`
minimax	MIT	`licenses/skills/LICENSE.minimax`
gemini-skills	Apache 2.0	`licenses/skills/LICENSE.gemini-skills`
ai-web3-security	MIT	`licenses/skills/LICENSE.ai-web3-security`
vuejs-nuxt	MIT	`licenses/skills/LICENSE.vuejs-nuxt`
ui-ux-pro-max	MIT	`licenses/skills/LICENSE.ui-ux-pro-max`

🙏 Support & Star

If StringRay helps you build better software, please consider:

⭐ Starring the repo on GitHub
📢 Sharing with your team
🐛 Reporting issues at github.com/htafolla/stringray/issues

📄 License

MIT License - see LICENSE for details.

Built with precision for enterprise-grade AI orchestration

Name		Name	Last commit message	Last commit date
Latest commit History 1,157 Commits
.githooks		.githooks
.github		.github
.husky		.husky
.opencode		.opencode
.strray		.strray
advanced-features		advanced-features
agents		agents
command		command
commands		commands
config		config
docs		docs
examples/file-monitoring		examples/file-monitoring
hooks		hooks
integrations		integrations
licenses/skills		licenses/skills
monitoring		monitoring
public		public
reports		reports
scripts		scripts
skills		skills
src		src
strray		strray
tasks		tasks
tests		tests
tweets		tweets
workflows		workflows
.gitignore		.gitignore
.npmignore		.npmignore
AGENTS-consumer.md		AGENTS-consumer.md
AGENTS-full.md		AGENTS-full.md
AGENTS.md		AGENTS.md
CHANGELOG-v1.15.x.md		CHANGELOG-v1.15.x.md
CHANGELOG.md		CHANGELOG.md
LICENSE		LICENSE
README.md		README.md
enforcer-config.json		enforcer-config.json
esbuild.json		esbuild.json
init.sh		init.sh
opencode.json		opencode.json
package-lock.json		package-lock.json
package.json		package.json
state		state
tsconfig.full.json		tsconfig.full.json
tsconfig.json		tsconfig.json

Folders and files

Latest commit

History

Repository files navigation

⚡ StringRay AI

What is StringRay?

Who is it for?

🚀 Quick Start

Standalone Mode (No OpenCode Required)

✨ Features

🤖 Available Agents

📦 OpenClaw Integration

What It Does

Quick Setup

🔮 Hermes Agent Integration

What You Get

Setup

Usage Examples

How It Differs from OpenCode

Tips

📖 Documentation

🔧 CLI Tools

⚙️ Configuration

Default Configuration

Customizing Agents

Feature Flags

Token Management

Version Pinning

📁 Project Structure

💬 Usage

🔌 Framework Integration

🎯 Skills

Skills Registry

Starter Packs

Registry Sources

Impeccable - AI Frontend Design

Framework Skills (30 Built-in)

License Information

🙏 Support & Star

📄 License

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases 6

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages