0xray
БесплатноНе проверенMulti-agent orchestration and Codex governance for OpenCode, Hermes, Grok Build, and OpenClaw
Описание
Multi-agent orchestration and Codex governance for OpenCode, Hermes, Grok Build, and OpenClaw
README
v3.5.4 — 42 agents · 45 skills · 7 MCP servers · 68 codex terms · 3,226 tests
Your AI Power Suit
Your AI Power ⚡️ Suit 🦾 — an exoskeleton that blocks AI slop from becoming real damage.
- learns with Repertoire 🧠
- has a Solar Ban Hammer ☀️🔨 (Dynamo governance)
- can also Confer 🗣️ (3-agent internal deliberation)
The core AI Governance OS stack is complete and open source. A commercial version is next — powers up the suit further.
npm install 0xray@latest
xray is the pure v2 three-subsystem AI orchestration framework — MCP-centric, governed by Dynamo, and autonomous via thinDispatch. Consumer npm install 0xray auto-wires all four platform bridges and seven MCP servers with zero manual config.
Quick Start
npm install 0xray # postinstall: 4 bridges + 7 MCPs + AGENTS.md + SKILLS.md + .mcp.json
npx 0xray status # verify install
npx 0xray setup # optional: symlinks, hook extras
Manual per-platform install (idempotent, same result as postinstall):
npx 0xray opencode install
npx 0xray grok install # 7 MCPs servers + dual skill sync (~/.grok/plugins + ~/.grok/skills)
npx 0xray hermes install
npx 0xray openclaw install
npx 0xray skill:install # starter skills
Autonomy command (default operating model)
When the suit is worn, autonomy-command is ON by default: lead dev takes helm, phased todos, best subagent per task, per-suite test triage, loop until green. No keywords required. Slash: /autonomy-command
Docs: guides/autonomy-command · Skill: src/skills/autonomy-command/SKILL.md
What's New Since 3.1
| Version | Highlights |
|---|---|
| 3.4.1 | Unified install-bridges.cjs on postinstall — OpenCode, Grok, Hermes, OpenClaw in one pass. All 7 MCPs servers via npx -y 0xray mcp <cmd> (no dist/ paths). Canonical release.mjs pipeline. |
| 3.3.1 | Orchestrator confidence gate wired into execution planning. |
| 3.3.0 | Pluggable Memory Routing (features.json → memory_routing). Repertoire is the default provider in the framework repo. |
| 3.2.0 | Typecheck hardening (58 errors fixed), orphan cleanup (5 deleted / ~39 integrated), full pre-tool-use hook, SelfProposalEngine + AsideContext restored, Hermes E2E 44/0/0, Grok CLI E2E green. |
| 3.1.1 | StringRay → 0xRay rename. Marketplace files (.mcp.json, .grok-plugin/plugin.json). Postinstall ships AGENTS.md, SKILLS.md, .gitignore.default. ConfigLoader mcpServers format. |
| 3.1.0 | Pre-publish guard tolerates missing features.json. |
Consolidations
- One consumer install path —
postinstall.cjs→installAllBridges()replaces scattered per-platform setup. - 7-server MCP surface —
.mcp.jsonSSOT; Grok plugin and all bridges shareXRAY_MCP_SERVERS. - Dev vs consumer AGENTS —
AGENTS.md(framework) vsAGENTS-consumer.md(copied to consumer projects on install). - Release pipeline —
npm run release:patch|minor|major→ reconcile → gate → artifacts → tag → publish.
Removals / Deprecations
- StringRay / strray-ai branding retired (3.1.1).
hermes bridgeCLI removed — usenpx 0xray hermes install..opencode/xray/fallback removed from auto-reflection-generator (3.1.1).- ~180 stale
@versionJSDoc tags and "xray 2.0" command doc strings cleaned (3.2.0). advanced-features/decoupled from core boot — not on the consumer install path.- PostProcessor soft-deprecated since 3.0 (
enablePostProcessor: falsedefault).
Three-Subsystem Architecture
┌─────────────────────────────────────────────────┐
│ Inference │
│ Proposals · Reflection · Memory routing │
├─────────────────────────────────────────────────┤
│ External Governance (Dynamo) │
│ Codex enforcement · Resonance/Isotopic · SSOT │
│ 3 deliberation MCPs: code-review, security, │
│ researcher (within 7-server consumer surface) │
├─────────────────────────────────────────────────┤
│ Autonomous Engine (thinDispatch) │
│ 7-flow · AsideContext · Confidence gate │
└─────────────────────────────────────────────────┘
- Inference — proposals, reflection, execution planning. Optional memory-routing enrichment (v3.3).
- External Governance — Dynamo Solar SSOT; CodexPolicyService; weighted PHI/TAU deliberation.
- Autonomous Engine — thinDispatch 7-flow routing; orchestrator confidence gate (v3.3.1).
Consumer Install (postinstall)
On npm install 0xray in a consumer project, postinstall automatically:
- Copies
AGENTS-consumer.md→AGENTS.mdand ships rootSKILLS.md - Seeds
.gitignorefrom.gitignore.default(if absent) - Deploys
.xray/config (codex.json,features.json,config.json) - Writes project
.mcp.jsonwith 7 MCPs servers (npx -y 0xray mcp …) - Installs 4 bridges: OpenCode (agents +
opencode.json), Grok (plugin + global skills), Hermes (~/.hermes/plugins/xray-hermes), OpenClaw (config + skills) - Syncs 45 framework skills to platform skill dirs
- Installs git pre-commit hook (non-blocking if not a git repo)
Seven MCP Servers (consumer)
All registered via npx -y 0xray mcp <cmd> — no brittle dist/ paths:
| Server | Command | Role |
|---|---|---|
xray-governance |
mcp governance |
Proposal governance, codex snapshot, quality gates |
xray-skills |
mcp skills |
45 knowledge skills + skill invocation |
xray-orchestrator |
mcp orchestrator |
Complexity analysis, lead-dev plan, task delegation |
xray-enforcer |
mcp enforcer |
Codex compliance, rule validation |
xray-researcher |
mcp researcher |
Codebase exploration, implementation lookup |
xray-code-review |
mcp code-review |
Proposal quality, best practices |
xray-architect-tools |
mcp architect-tools |
System design, architecture decisions |
The framework repo also ships additional internal .server.ts implementations for orchestration pipelines — these are not part of the 7-server consumer .mcp.json surface.
Agents
42 YML agent surfaces in src/opencode/agents/*.yml — zero manual registration. Invoke via @agent-name in OpenCode.
Core governance agents:
| Agent | Purpose |
|---|---|
@enforcer |
Codex compliance & error prevention |
@orchestrator |
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 |
@bug-triage-specialist |
Error investigation |
@researcher |
Codebase exploration |
See AGENTS.md and SKILLS.md for the full agent and skill catalog.
Memory Routing + Repertoire (v3.3+)
Pluggable memory_routing block in features.json (validated by features.schema.json):
"memory_routing": {
"enabled": true,
"provider": "repertoire",
"module_path": "../repertoire/dist/provider/memory-routing-provider.js",
"config": {
"dataDir": "../repertoire/data",
"signalsPath": "../repertoire/data/curated_signals.json",
"logDir": "../repertoire/logs/groover-inference"
}
}
| Integration | What it does |
|---|---|
| ExecutionPlanner (v3.3.1) | getTaskConfidence() → complexity boost, trap hints, signal-aware selectAgent() |
| thinDispatch | Nucleus scoreAndRoute() (0xray/nucleus); memory provider resolveThinDispatch() when routing enabled |
| Researcher | researcher-confidence.ts appends MEMORY_ROUTING: block to governance output |
| AsideContext (v3.2) | buildInheritedContext → inheritedContext.memoryRouting on orchestrator spawnAside |
| Feedback | Per-task ingestFeedback() closes the learning loop |
External hosts (Hermes/Grok): add repertoire-mcp to .mcp.json — see Repertoire docs.
Consumers without Repertoire: "memory_routing": { "enabled": false, "provider": "null" }.
Docs: memory routing · Repertoire · features.json · all features since 3.1
Integrations
| Platform | Install | Postinstall behavior |
|---|---|---|
| OpenCode | npx 0xray opencode install |
Merges opencode.json, copies agent YML surfaces |
| Grok CLI / Build | npx 0xray grok install |
Plugin + ~/.grok/skills/ sync, 7 MCPs servers |
| Hermes Agent | npx 0xray hermes install |
~/.hermes/plugins/xray-hermes, consumer root marker |
| OpenClaw | npx 0xray openclaw install |
.xray/config/openclaw.json, skill sync |
Governance & Codex
- 68 terms in
.xray/codex.json— core, architecture, testing, performance, security, operations, governance - CodexPolicyService — Governance-owned SSOT for codex loading
- Pre-governance gate blocks non-compliant proposals
- Active codex snapshot via
get_active_codexMCP tool - frameworkLogger structured logging (never
console.*)
Testing
| Suite | Status (v3.4.1) |
|---|---|
| Unit / Integration | 185 files, 3,226 passed |
| OpenCode E2E | 42/42 solo, 34/34 orchestrator |
| Grok CLI E2E | 62/0 failures (v3.2.0 verified) |
| OpenClaw E2E | 9/9 (v3.2.0 verified) |
| Hermes E2E | 44/0/0 (v3.2.0 verified) |
| Consumer smoke | npm run release:gate — pack → clean install → 7 MCPs + 4 bridges |
npm test
npm run release:gate # full release gate
Release
npm run release:patch # canonical pipeline
npm run release:minor
npm run release:major
npm run release:patch -- --dry-run
Pipeline: reconcile-version → release-gate (build + test + consumer smoke) → CHANGELOG/README/AGENTS artifacts → commit → tag → npm publish.
License
MIT — see LICENSE
xray — MCP-centric, governed, autonomous. Pure v2 three-subsystem.
Установить 0xray в Claude Desktop, Claude Code, Cursor
unyly install 0xrayСтавит в Claude Desktop, Claude Code, Cursor и VS Code — сам разбирается с npx, uvx и сборкой из исходников.
Впервые? Поставь CLI: curl -fsSL https://unyly.org/install | sh
Или настроить вручную
Выполни в терминале:
claude mcp add 0xray -- npx -y 0xrayFAQ
0xray MCP бесплатный?
Да, 0xray MCP бесплатный — установка в пару кликов через Unyly без оплаты.
Нужен ли API-ключ для 0xray?
Нет, 0xray работает без API-ключей и переменных окружения.
0xray — hosted или self-hosted?
Self-hosted: сервер запускается локально на твоей машине командой из раздела установки.
Как установить 0xray в Claude Desktop, Claude Code или Cursor?
Открой 0xray на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.
Похожие MCP
Fetch
Web content fetching and conversion for efficient LLM usage.
AWS KB Retrieval
Retrieval from AWS Knowledge Base using Bedrock Agent Runtime.
автор: modelcontextprotocolSpring AI MCP Server
Provides auto-configuration for setting up an MCP server in Spring Boot applications.
llm-analysis-assistant
A very streamlined mcp client that supports calling and monitoring stdio/sse/streamableHttp, and can also view request responses through the /logs page. It also
автор: xuzexin-hzCompare 0xray with
Не уверен что выбрать?
Найди свой стек за 60 секунд
Автор?
Embed-бейдж для README
Похожее
Все в категории ai
