[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77TYZ9SHZFTB0NFKC64WBQY":3,"skill-same:01M77TYZ9SHZFTB0NFKC64WBQY":47},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":6,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M77TYZ9SHZFTB0NFKC64WBQY","multi-agent-deployment",null,"Deploy production-grade multi-agent fleets in OpenClaw with battle-tested scripts, cloud deployment templates, and shared memory infrastructure. Use when: (1) Deploying multiple specialized AI agents with routing, (2) Setting up shared memory between agents, (3) Deploying to cloud platforms (DigitalOcean, AWS, GCP, K8s), (4) Building agent teams for business workflows, (5) Moving from single-agent to production multi-agent setups. Includes working Python scripts, cloud deployment configs, and troubleshooting guides based on real deployments.","Lord1Egypt\u002Fawesome-skill-forge","Lord1Egypt","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F99746359?v=4","01M7592YDNT0XK9WJ0T2MQVPDN","Mohamed Mounir Ali AL Keshita","01M75GRFNHC8VZNWEB8DKDE433",3,"MIT","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fs\u002Fskill-multi-agent-deployment\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: multi-agent-deployment\ndescription: >\n  Deploy production-grade multi-agent fleets in OpenClaw with battle-tested\n  scripts, cloud deployment templates, and shared memory infrastructure.\n  Use when: (1) Deploying multiple specialized AI agents with routing,\n  (2) Setting up shared memory between agents, (3) Deploying to cloud\n  platforms (DigitalOcean, AWS, GCP, K8s), (4) Building agent teams for\n  business workflows, (5) Moving from single-agent to production multi-agent\n  setups. Includes working Python scripts, cloud deployment configs, and\n  troubleshooting guides based on real deployments.\n---\n\n# Multi-Agent Deployment Skill for OpenClaw\n\n**Production deployment focus.** This skill provides **executable scripts** and\n**cloud-ready configurations** — not just documentation. Unlike pure\ncoordination-pattern skills, this gives you working infrastructure:\n`agent_setup.py` creates your agent directories, `routing_config.py`\ngenerates OpenClaw config, `memory_sync.py` wires shared memory,\n`fleet_validate.py` checks your deployment health, and\n`deploy_script.sh` ships to DigitalOcean, AWS, GCP, Azure, or Kubernetes.\n\n| Who this is for | What you get |\n|---|---|\n| OpenClaw users running 1 agent → need a fleet | Scripts, templates, working configs |\n| Developers deploying to cloud (DO\u002FAWS\u002FGCP) | Platform YAML\u002FJSON templates + Dockerfile |\n| Teams needing shared memory across agents | File-based shared memory with locking + TTL |\n| Anyone testing multi-agent architectures | 5 agent type templates + routing rules |\n\n---\n\n## Quick Start (5 minutes)\n\n```bash\n# 1. Create 5 agent directories with SOUL.md + AGENTS.md templates\npython scripts\u002Fagent_setup.py --agents coordinator,research,builder,auditor,personal --output .\u002Fagents\n\n# 2. Generate OpenClaw routing config\npython scripts\u002Frouting_config.py --input .\u002Fagents --output ~\u002F.openclaw\u002Fconfig.json\n\n# 3. Initialize shared memory\npython scripts\u002Fmemory_sync.py --init --path .\u002Fshared_memory\n\n# 4. Validate your fleet (new!)\npython scripts\u002Ffleet_validate.py --agents .\u002Fagents --config ~\u002F.openclaw\u002Fconfig.json\n\n# 5. Deploy (or dry-run first)\nbash scripts\u002Fdeploy_script.sh --platform digitalocean --region nyc3 --dry-run\n```\n\n---\n\n## What's Included\n\n### 🐍 Python Scripts (working, tested)\n\n| Script | Purpose | Key Features |\n|--------|---------|-------------|\n| `scripts\u002Fagent_setup.py` | Scaffold agent directories | Reads templates from `assets\u002Ftemplates\u002F`, creates 5 agent types + shared dir |\n| `scripts\u002Frouting_config.py` | Generate OpenClaw config.json | Per-agent routing rules, fallback to coordinator, shared memory wiring, **health probes**, **rate limits** |\n| `scripts\u002Fmemory_sync.py` | File-based shared memory | JSON persistence, file-locking, TTL per key, event journaling, agent stats, **typed schema validation** |\n| `scripts\u002Ffleet_validate.py` | Validate deployment health | Agent structure, routing config, shared memory, template integrity check |\n| `scripts\u002Fdeploy_script.sh` | One-click cloud deploy | DO App Spec, AWS ECS, GCP Cloud Run, Azure ACI, K8s manifests |\n\n### 📁 Asset Templates\n| File | Purpose |\n|------|--------|\n| `assets\u002Ftemplates\u002FSOUL.md` | Agent personality + mission template (populated by agent_setup.py) |\n| `assets\u002Ftemplates\u002FAGENTS.md` | Operating rules and protocols template |\n\n### 📚 Reference Docs\n- [references\u002Farchitecture.md](references\u002Farchitecture.md): Full agent architecture patterns (coordinator-centric, peer-to-peer, pipeline)\n- [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md): 6 common failure patterns with recovery procedures\n- [references\u002Fworkflows.md](references\u002Fworkflows.md): Workflow DAG templates with pipeline execution patterns (v1.2.0+)\n- [references\u002Fmemory-schemas.md](references\u002Fmemory-schemas.md): Typed inter-agent memory schemas with validation rules (v1.2.0+)\n\n### ☁️ Cloud Deployment Configs\n- `deployments\u002Fdigitalocean\u002Fapp.yaml` — DigitalOcean App Platform\n- `deployments\u002Faws\u002Ftask-definition.json` — AWS ECS\n- `deployments\u002Fgcp\u002Fcloudrun.yaml` — GCP Cloud Run\n- `deployments\u002Fazure\u002Fcontainer-instance.yaml` — Azure Container Instances\n- `deployments\u002Fk8s\u002Fdeployment.yaml` — Kubernetes deployment + HPA\n- `deployments\u002FDockerfile.openclaw` — Container image for gateway\n*(Shipped with skill — customize or regenerate with deploy_script.sh)*\n\n---\n\n## Architecture Overview\n\nThe skill supports three deployment-tested agent topologies:\n\n### 1. Coordinator-Centric (recommended start)\n```\nUser → Coordinator → [Research, Builder, Auditor, Personal]\n                      → Shared Memory ← all\n```\nBest for: business workflows, clear task handoffs, teams \u003C 10 agents.\n\n### 2. Peer-to-Peer\n```\nAgents ↔ Event Bus ↔ Shared Memory\n```\nBest for: creative\u002Fresearch workflows, emergent coordination, high collaboration.\n\n### 3. Pipeline\n```\nResearch → Builder → Auditor → Deploy\n```\nBest for: content creation, data processing, linear workflows.\n\n*See [references\u002Farchitecture.md](references\u002Farchitecture.md) for full topology diagrams and config examples.*\n\n---\n\n## Shared Memory Design\n\nTwo backends available — **file** (default, zero deps) and **SQLite** (production).\n\n**File backend** (zero infrastructure):\n- **Atomic writes** via temp file + replace\n- **File locking** with configurable timeout\n- **TTL-based expiration** on shared keys\n- **Event journaling** (daily JSONL files)\n- **Agent statistics** (read\u002Fwrite counts, last active)\n- **Auto-cleanup** of expired entries\n- **Typed schema validation** for structured inter-agent data (v1.2.0+)\n\n```bash\n# Write data (free-form, file backend)\npython scripts\u002Fmemory_sync.py --write 'coordinator:task_status:{\"status\":\"active\"}'\n\n# Write data with schema validation\npython scripts\u002Fmemory_sync.py --write 'coordinator:current_task:{\"task_id\":\"t001\",\"assigned_by\":\"coordinator\",\"assigned_to\":\"research\",\"priority\":\"high\",\"title\":\"Research trends\",\"description\":\"Analyze marketplace\"}' --schema task_assignment\n\n# Read data\npython scripts\u002Fmemory_sync.py --read 'research:search_results'\n\n# View agent stats\npython scripts\u002Fmemory_sync.py --stats\n\n# List available typed schemas\npython scripts\u002Fmemory_sync.py --list-schemas\n\n# List all shared keys (v1.3.0)\npython scripts\u002Fmemory_sync.py --keys\n\n# Count keys per agent (v1.3.0)\npython scripts\u002Fmemory_sync.py --count-by-agent\n\n# Cleanup expired\npython scripts\u002Fmemory_sync.py --cleanup\n\n# === SQLite Backend (production, v1.3.0) ===\n\n# Initialize SQLite shared memory\npython scripts\u002Fmemory_sync.py --backend sqlite --init --path .\u002Fshared_memory\n\n# Write with SQLite backend\npython scripts\u002Fmemory_sync.py --backend sqlite --write 'coordinator:task:{\"id\":\"t1\"}'\n\n# Migrate from file to SQLite\npython scripts\u002Fmemory_sync.py --migrate-to-sqlite .\u002Fshared_memory\n\n# Reclaim SQLite space\npython scripts\u002Fmemory_sync.py --backend sqlite --compact\n```\n\nFor high-concurrency production, use the **SQLite backend** (`--backend sqlite`) with WAL mode\nfor concurrent reads\u002Fwrites and ACID transactions — all from the Python std library, no pip install needed.\n\nMigrate from file to SQLite:\n```bash\npython scripts\u002Fmemory_sync.py --migrate-to-sqlite .\u002Fshared_memory\n```\n\n### HTTP REST API (v1.4.0+)\n\nExpose shared memory over HTTP using **only Python stdlib** — no pip install needed.\nStart the API server with `--listen`:\n\n```bash\n# Start server (default: port 8080, bind 127.0.0.1)\npython scripts\u002Fmemory_sync.py --listen --backend sqlite --path .\u002Fshared_memory\n```\n\n| Method | Path | Description |\n|--------|------|-------------|\n| GET | `\u002Fhealth` | Health check |\n| GET | `\u002Fkeys?agent=X` | List keys (optional agent filter) |\n| GET | `\u002Fcount` | Keys per agent |\n| GET | `\u002Fstats` | Full statistics |\n| GET | `\u002Fschemas` | Available typed schemas |\n| GET | `\u002Fread\u002F\u003Ca>\u002F\u003Ck>` | Read a value |\n| POST | `\u002Fwrite` | Write a value (JSON body) |\n\n```bash\n# Write via API\ncurl -X POST http:\u002F\u002Flocalhost:8080\u002Fwrite -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"agent_id\":\"coordinator\",\"key\":\"task:t001\",\"value\":{\"task_id\":\"t001\",\"assigned_to\":\"research\"},\"schema\":\"task_assignment\"}'\n```\n\nCORS headers included for web-based dashboards.\n\n---\n\n## Cloud Deployment\n\n```bash\n# Dry-run (generates configs only)\nbash scripts\u002Fdeploy_script.sh --platform aws --region us-east-1 --dry-run\n\n# Actual deploy\nbash scripts\u002Fdeploy_script.sh --platform digitalocean --region nyc3 --size s-2vcpu-2gb\n```\n\nThe script:\n1. Detects or accepts platform (`--platform digitalocean|aws|gcp|azure|k8s`)\n2. Generates platform-specific configs under `deployments\u002F\u003Cplatform>\u002F`\n3. Dockerfile included for containerized gateway\n4. Generates deployment summary in `deployments\u002Fdeployment-summary.md`\n\n---\n\n## Common Use Cases\n\n### E-commerce Support Team\nCustomer → Coordinator → [Support, Inventory, Analytics, Order]\n\n### Content Creation Team\nResearch → Writer → Editor → Publisher (pipeline pattern)\n\n### Development Team\nPlanner → Coder → Tester → Deployer → Auditor (pipeline + audit gate)\n\n### Personal Productivity\nCoordinator → [Research, Builder, Personal Assistant]\n\n---\n\n## Agent Types Included\n\n| Agent | Mission | Default Router |\n|-------|---------|---------------|\n| **coordinator** | Route requests, monitor availability, resolve conflicts | `help`, `assist`, `support` |\n| **research** | Gather and synthesize web information | `research`, `find`, `search`, `look up`, `what is` |\n| **builder** | Write code, create files, debug | `build`, `create`, `write`, `code`, `script`, `develop` |\n| **auditor** | Review quality, security, compliance | `review`, `check`, `audit`, `validate`, `test` |\n| **personal** | Scheduling, reminders, communication | `schedule`, `remind`, `calendar`, `email`, `message` |\n\n---\n\n## Monitoring, Security & Recovery\n\n### Agent Monitoring\n- Shared memory stats: `python scripts\u002Fmemory_sync.py --stats`\n- SQLite stats (production): `python scripts\u002Fmemory_sync.py --backend sqlite --stats`\n- List all keys: `python scripts\u002Fmemory_sync.py --keys`\n- Keys per agent: `python scripts\u002Fmemory_sync.py --count-by-agent`\n- Event journal: `tail -f shared_memory\u002Fevents\u002F$(date +%Y%m%d).jsonl`\n- System health: `openclaw agents list && openclaw gateway status`\n\n### Security Perimeter\n| Layer | Protection |\n|-------|-----------|\n| Workspace isolation | Each agent has separate dirs |\n| Memory isolation | Shared memory via API only (file-locked) |\n| Tool restrictions | Only necessary tools per agent type |\n| Audit logging | All writes logged to event journal |\n\n### Recovery Procedures\nSee [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md) for:\n- Complete system failure recovery\n- Data corruption restoration\n- Individual agent recovery\n- Configuration drift fixes\n\n---\n\n## Why This Skill Over Competitors\n\n| Dimension | This Skill | Pure Coordination Skills |\n|-----------|-----------|------------------------|\n| **Focus** | Production deployment | Agent coordination patterns |\n| **Scripts** | 4 working Python\u002Fbash scripts | Usually documentation-only |\n| **Cloud** | Multi-platform (DO\u002FAWS\u002FGCP\u002FAzure\u002FK8s) | Rarely included |\n| **Templates** | SOUL.md + AGENTS.md (populated at runtime) | Often static examples |\n| **Shared Memory** | Working file-based system with locking, TTL, stats | Conceptual description |\n| **Troubleshooting** | 6 failure patterns with executable recovery steps | Generic advice |\n\n---\n\n## OpenClaw 2026.x Integration Patterns\n\nThis skill's outputs work with OpenClaw 2026.2+ and 2026.3+ features:\n- `sessions_spawn` for sub-agent task delegation\n- `cron` for scheduled multi-agent workflows\n- Shared memory as communication bridge between agents\n- Routing rules for message-to-agent direction\n- `fleet_validate.py` for post-deployment health checks\n- `openclaw agents` management commands for fleet lifecycle\n\n### 2026.3+ Feature Notes\n\nThe latest routing config generator (`routing_config.py`) automatically generates\nconfigurations for these 2026.3+ features—no manual editing needed:\n\n- **Agent health probes**: Each agent entry has `healthCheck` config\n  with interval, timeout, and retry settings\n- **Gateway-level rate limiting**: Per-agent `rateLimit` settings\n  (30 req\u002Fmin, 5 concurrent)\n- **Structured memory types**: `memory_sync.py` now supports typed schema\n  validation via `--schema` flag. See\n  [references\u002Fmemory-schemas.md](references\u002Fmemory-schemas.md) for\n  all available schemas (task_assignment, status_update, research_finding,\n  etc.)\n- **Workflow DAGs**: 5 ready-to-use pipeline execution templates in\n  [references\u002Fworkflows.md](references\u002Fworkflows.md) — content creation,\n  development sprints, customer support, data pipelines, and daily briefings\n  with topological dependency ordering and execution scripts\n\n---\n\n## Troubleshooting Quick Reference\n\n**Agents not responding?**\n```bash\nopenclaw agents list           # Are they running?\njournalctl -u openclaw-gateway | tail -20  # Any errors?\n```\n\n**Wrong routing?**\n```bash\npython scripts\u002Frouting_config.py --input .\u002Fagents --config-output .\u002Frouting-test.json\n```\n\n**Shared memory corruption?**\n```bash\npython -m json.tool shared_memory\u002Fshared_data.json  # Validate JSON\npython scripts\u002Fmemory_sync.py --cleanup              # Remove expired\npython scripts\u002Fmemory_sync.py --list-schemas         # See available typed schemas\n```\n\n**Fleet health check?**\n```bash\npython scripts\u002Ffleet_validate.py --agents .\u002Fagents --config ~\u002F.openclaw\u002Fconfig.json\n```\n\n**Schema validation on write?**\n```bash\npython scripts\u002Fmemory_sync.py --write 'coordinator:task:{\"task_id\":\"t1\",\"assigned_to\":\"research\"}' --schema task_assignment\n```\n\n---\n\n## Version History\n\n| Version | Date | Changes |\n|---------|------|--------|\n| 1.4.0 | 2026-05-27 | **HTTP REST API** — `--listen` flag starts zero-dep REST server exposing shared memory over HTTP. GET\u002FPOST for read, write, keys, stats, count, cleanup, schemas, health. CORS headers, stdlib only |\n| 1.3.0 | 2026-05-26 | **SQLite backend** — WAL-mode concurrent shared memory, ACID transactions, `--backend sqlite`, `--migrate-to-sqlite` tool, `--keys` and `--count-by-agent` utilities, `--compact` for space reclamation. No pip deps required. SKILL.md and memory-schemas.md updated with SQLite docs |\n| 1.2.0 | 2026-05-25 | Structured memory schemas with validation; health probes in routing config; memory-schemas.md reference; rate limiting; `--schema` and `--list-schemas` flags; **workflows.md** with 5 DAG templates; schema validation bug fix (inner data fields + correct schema type); reference list expanded |\n| 1.1.0 | 2026-05-23 | Fleet validation script added; deployment config templates shipped; date bug fixed; OpenClaw 2026.3+ pattern notes |\n| 1.0.1 | 2026-05-11 | Templates unified; agent_setup.py reads from assets\u002Ftemplates\u002F; _meta.json + origin tracking added |\n| 1.0.0 | 2026-04-02 | Initial release — agent setup, routing config, memory sync, cloud deploy |\n\n---\n\n*Based on production patterns from Abhi's multi-agent deployment experience. Tested with OpenClaw 2026.2+.*\n*Published on [ClawHub](https:\u002F\u002Fclawhub.ai) — slug: multi-agent-deployment*","# Multi-Agent Deployment Skill for OpenClaw\n\n**Production deployment focus.** This skill provides **executable scripts** and\n**cloud-ready configurations** — not just documentation. Unlike pure\ncoordination-pattern skills, this gives you working infrastructure:\n`agent_setup.py` creates your agent directories, `routing_config.py`\ngenerates OpenClaw config, `memory_sync.py` wires shared memory,\n`fleet_validate.py` checks your deployment health, and\n`deploy_script.sh` ships to DigitalOcean, AWS, GCP, Azure, or Kubernetes.\n\n| Who this is for | What you get |\n|---|---|\n| OpenClaw users running 1 agent → need a fleet | Scripts, templates, working configs |\n| Developers deploying to cloud (DO\u002FAWS\u002FGCP) | Platform YAML\u002FJSON templates + Dockerfile |\n| Teams needing shared memory across agents | File-based shared memory with locking + TTL |\n| Anyone testing multi-agent architectures | 5 agent type templates + routing rules |\n\n---\n\n## Quick Start (5 minutes)\n\n```bash\n# 1. Create 5 agent directories with SOUL.md + AGENTS.md templates\npython scripts\u002Fagent_setup.py --agents coordinator,research,builder,auditor,personal --output .\u002Fagents\n\n# 2. Generate OpenClaw routing config\npython scripts\u002Frouting_config.py --input .\u002Fagents --output ~\u002F.openclaw\u002Fconfig.json\n\n# 3. Initialize shared memory\npython scripts\u002Fmemory_sync.py --init --path .\u002Fshared_memory\n\n# 4. Validate your fleet (new!)\npython scripts\u002Ffleet_validate.py --agents .\u002Fagents --config ~\u002F.openclaw\u002Fconfig.json\n\n# 5. Deploy (or dry-run first)\nbash scripts\u002Fdeploy_script.sh --platform digitalocean --region nyc3 --dry-run\n```\n\n---\n\n## What's Included\n\n### 🐍 Python Scripts (working, tested)\n\n| Script | Purpose | Key Features |\n|--------|---------|-------------|\n| `scripts\u002Fagent_setup.py` | Scaffold agent directories | Reads templates from `assets\u002Ftemplates\u002F`, creates 5 agent types + shared dir |\n| `scripts\u002Frouting_config.py` | Generate OpenClaw config.json | Per-agent routing rules, fallback to coordinator, shared memory wiring, **health probes**, **rate limits** |\n| `scripts\u002Fmemory_sync.py` | File-based shared memory | JSON persistence, file-locking, TTL per key, event journaling, agent stats, **typed schema validation** |\n| `scripts\u002Ffleet_validate.py` | Validate deployment health | Agent structure, routing config, shared memory, template integrity check |\n| `scripts\u002Fdeploy_script.sh` | One-click cloud deploy | DO App Spec, AWS ECS, GCP Cloud Run, Azure ACI, K8s manifests |\n\n### 📁 Asset Templates\n| File | Purpose |\n|------|--------|\n| `assets\u002Ftemplates\u002FSOUL.md` | Agent personality + mission template (populated by agent_setup.py) |\n| `assets\u002Ftemplates\u002FAGENTS.md` | Operating rules and protocols template |\n\n### 📚 Reference Docs\n- [references\u002Farchitecture.md](references\u002Farchitecture.md): Full agent architecture patterns (coordinator-centric, peer-to-peer, pipeline)\n- [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md): 6 common failure patterns with recovery procedures\n- [references\u002Fworkflows.md](references\u002Fworkflows.md): Workflow DAG templates with pipeline execution patterns (v1.2.0+)\n- [references\u002Fmemory-schemas.md](references\u002Fmemory-schemas.md): Typed inter-agent memory schemas with validation rules (v1.2.0+)\n\n### ☁️ Cloud Deployment Configs\n- `deployments\u002Fdigitalocean\u002Fapp.yaml` — DigitalOcean App Platform\n- `deployments\u002Faws\u002Ftask-definition.json` — AWS ECS\n- `deployments\u002Fgcp\u002Fcloudrun.yaml` — GCP Cloud Run\n- `deployments\u002Fazure\u002Fcontainer-instance.yaml` — Azure Container Instances\n- `deployments\u002Fk8s\u002Fdeployment.yaml` — Kubernetes deployment + HPA\n- `deployments\u002FDockerfile.openclaw` — Container image for gateway\n*(Shipped with skill — customize or regenerate with deploy_script.sh)*\n\n---\n\n## Architecture Overview\n\nThe skill supports three deployment-tested agent topologies:\n\n### 1. Coordinator-Centric (recommended start)\n```\nUser → Coordinator → [Research, Builder, Auditor, Personal]\n                      → Shared Memory ← all\n```\nBest for: business workflows, clear task handoffs, teams \u003C 10 agents.\n\n### 2. Peer-to-Peer\n```\nAgents ↔ Event Bus ↔ Shared Memory\n```\nBest for: creative\u002Fresearch workflows, emergent coordination, high collaboration.\n\n### 3. Pipeline\n```\nResearch → Builder → Auditor → Deploy\n```\nBest for: content creation, data processing, linear workflows.\n\n*See [references\u002Farchitecture.md](references\u002Farchitecture.md) for full topology diagrams and config examples.*\n\n---\n\n## Shared Memory Design\n\nTwo backends available — **file** (default, zero deps) and **SQLite** (production).\n\n**File backend** (zero infrastructure):\n- **Atomic writes** via temp file + replace\n- **File locking** with configurable timeout\n- **TTL-based expiration** on shared keys\n- **Event journaling** (daily JSONL files)\n- **Agent statistics** (read\u002Fwrite counts, last active)\n- **Auto-cleanup** of expired entries\n- **Typed schema validation** for structured inter-agent data (v1.2.0+)\n\n```bash\n# Write data (free-form, file backend)\npython scripts\u002Fmemory_sync.py --write 'coordinator:task_status:{\"status\":\"active\"}'\n\n# Write data with schema validation\npython scripts\u002Fmemory_sync.py --write 'coordinator:current_task:{\"task_id\":\"t001\",\"assigned_by\":\"coordinator\",\"assigned_to\":\"research\",\"priority\":\"high\",\"title\":\"Research trends\",\"description\":\"Analyze marketplace\"}' --schema task_assignment\n\n# Read data\npython scripts\u002Fmemory_sync.py --read 'research:search_results'\n\n# View agent stats\npython scripts\u002Fmemory_sync.py --stats\n\n# List available typed schemas\npython scripts\u002Fmemory_sync.py --list-schemas\n\n# List all shared keys (v1.3.0)\npython scripts\u002Fmemory_sync.py --keys\n\n# Count keys per agent (v1.3.0)\npython scripts\u002Fmemory_sync.py --count-by-agent\n\n# Cleanup expired\npython scripts\u002Fmemory_sync.py --cleanup\n\n# === SQLite Backend (production, v1.3.0) ===\n\n# Initialize SQLite shared memory\npython scripts\u002Fmemory_sync.py --backend sqlite --init --path .\u002Fshared_memory\n\n# Write with SQLite backend\npython scripts\u002Fmemory_sync.py --backend sqlite --write 'coordinator:task:{\"id\":\"t1\"}'\n\n# Migrate from file to SQLite\npython scripts\u002Fmemory_sync.py --migrate-to-sqlite .\u002Fshared_memory\n\n# Reclaim SQLite space\npython scripts\u002Fmemory_sync.py --backend sqlite --compact\n```\n\nFor high-concurrency production, use the **SQLite backend** (`--backend sqlite`) with WAL mode\nfor concurrent reads\u002Fwrites and ACID transactions — all from the Python std library, no pip install needed.\n\nMigrate from file to SQLite:\n```bash\npython scripts\u002Fmemory_sync.py --migrate-to-sqlite .\u002Fshared_memory\n```\n\n### HTTP REST API (v1.4.0+)\n\nExpose shared memory over HTTP using **only Python stdlib** — no pip install needed.\nStart the API server with `--listen`:\n\n```bash\n# Start server (default: port 8080, bind 127.0.0.1)\npython scripts\u002Fmemory_sync.py --listen --backend sqlite --path .\u002Fshared_memory\n```\n\n| Method | Path | Description |\n|--------|------|-------------|\n| GET | `\u002Fhealth` | Health check |\n| GET | `\u002Fkeys?agent=X` | List keys (optional agent filter) |\n| GET | `\u002Fcount` | Keys per agent |\n| GET | `\u002Fstats` | Full statistics |\n| GET | `\u002Fschemas` | Available typed schemas |\n| GET | `\u002Fread\u002F\u003Ca>\u002F\u003Ck>` | Read a value |\n| POST | `\u002Fwrite` | Write a value (JSON body) |\n\n```bash\n# Write via API\ncurl -X POST http:\u002F\u002Flocalhost:8080\u002Fwrite -H \"Content-Type: application\u002Fjson\" \\\n  -d '{\"agent_id\":\"coordinator\",\"key\":\"task:t001\",\"value\":{\"task_id\":\"t001\",\"assigned_to\":\"research\"},\"schema\":\"task_assignment\"}'\n```\n\nCORS headers included for web-based dashboards.\n\n---\n\n## Cloud Deployment\n\n```bash\n# Dry-run (generates configs only)\nbash scripts\u002Fdeploy_script.sh --platform aws --region us-east-1 --dry-run\n\n# Actual deploy\nbash scripts\u002Fdeploy_script.sh --platform digitalocean --region nyc3 --size s-2vcpu-2gb\n```\n\nThe script:\n1. Detects or accepts platform (`--platform digitalocean|aws|gcp|azure|k8s`)\n2. Generates platform-specific configs under `deployments\u002F\u003Cplatform>\u002F`\n3. Dockerfile included for containerized gateway\n4. Generates deployment summary in `deployments\u002Fdeployment-summary.md`\n\n---\n\n## Common Use Cases\n\n### E-commerce Support Team\nCustomer → Coordinator → [Support, Inventory, Analytics, Order]\n\n### Content Creation Team\nResearch → Writer → Editor → Publisher (pipeline pattern)\n\n### Development Team\nPlanner → Coder → Tester → Deployer → Auditor (pipeline + audit gate)\n\n### Personal Productivity\nCoordinator → [Research, Builder, Personal Assistant]\n\n---\n\n## Agent Types Included\n\n| Agent | Mission | Default Router |\n|-------|---------|---------------|\n| **coordinator** | Route requests, monitor availability, resolve conflicts | `help`, `assist`, `support` |\n| **research** | Gather and synthesize web information | `research`, `find`, `search`, `look up`, `what is` |\n| **builder** | Write code, create files, debug | `build`, `create`, `write`, `code`, `script`, `develop` |\n| **auditor** | Review quality, security, compliance | `review`, `check`, `audit`, `validate`, `test` |\n| **personal** | Scheduling, reminders, communication | `schedule`, `remind`, `calendar`, `email`, `message` |\n\n---\n\n## Monitoring, Security & Recovery\n\n### Agent Monitoring\n- Shared memory stats: `python scripts\u002Fmemory_sync.py --stats`\n- SQLite stats (production): `python scripts\u002Fmemory_sync.py --backend sqlite --stats`\n- List all keys: `python scripts\u002Fmemory_sync.py --keys`\n- Keys per agent: `python scripts\u002Fmemory_sync.py --count-by-agent`\n- Event journal: `tail -f shared_memory\u002Fevents\u002F$(date +%Y%m%d).jsonl`\n- System health: `openclaw agents list && openclaw gateway status`\n\n### Security Perimeter\n| Layer | Protection |\n|-------|-----------|\n| Workspace isolation | Each agent has separate dirs |\n| Memory isolation | Shared memory via API only (file-locked) |\n| Tool restrictions | Only necessary tools per agent type |\n| Audit logging | All writes logged to event journal |\n\n### Recovery Procedures\nSee [references\u002Ftroubleshooting.md](references\u002Ftroubleshooting.md) for:\n- Complete system failure recovery\n- Data corruption restoration\n- Individual agent recovery\n- Configuration drift fixes\n\n---\n\n## Why This Skill Over Competitors\n\n| Dimension | This Skill | Pure Coordination Skills |\n|-----------|-----------|------------------------|\n| **Focus** | Production deployment | Agent coordination patterns |\n| **Scripts** | 4 working Python\u002Fbash scripts | Usually documentation-only |\n| **Cloud** | Multi-platform (DO\u002FAWS\u002FGCP\u002FAzure\u002FK8s) | Rarely included |\n| **Templates** | SOUL.md + AGENTS.md (populated at runtime) | Often static examples |\n| **Shared Memory** | Working file-based system with locking, TTL, stats | Conceptual description |\n| **Troubleshooting** | 6 failure patterns with executable recovery steps | Generic advice |\n\n---\n\n## OpenClaw 2026.x Integration Patterns\n\nThis skill's outputs work with OpenClaw 2026.2+ and 2026.3+ features:\n- `sessions_spawn` for sub-agent task delegation\n- `cron` for scheduled multi-agent workflows\n- Shared memory as communication bridge between agents\n- Routing rules for message-to-agent direction\n- `fleet_validate.py` for post-deployment health checks\n- `openclaw agents` management commands for fleet lifecycle\n\n### 2026.3+ Feature Notes\n\nThe latest routing config generator (`routing_config.py`) automatically generates\nconfigurations for these 2026.3+ features—no manual editing needed:\n\n- **Agent health probes**: Each agent entry has `healthCheck` config\n  with interval, timeout, and retry settings\n- **Gateway-level rate limiting**: Per-agent `rateLimit` settings\n  (30 req\u002Fmin, 5 concurrent)\n- **Structured memory types**: `memory_sync.py` now supports typed schema\n  validation via `--schema` flag. See\n  [references\u002Fmemory-schemas.md](references\u002Fmemory-schemas.md) for\n  all available schemas (task_assignment, status_update, research_finding,\n  etc.)\n- **Workflow DAGs**: 5 ready-to-use pipeline execution templates in\n  [references\u002Fworkflows.md](references\u002Fworkflows.md) — content creation,\n  development sprints, customer support, data pipelines, and daily briefings\n  with topological dependency ordering and execution scripts\n\n---\n\n## Troubleshooting Quick Reference\n\n**Agents not responding?**\n```bash\nopenclaw agents list           # Are they running?\njournalctl -u openclaw-gateway | tail -20  # Any errors?\n```\n\n**Wrong routing?**\n```bash\npython scripts\u002Frouting_config.py --input .\u002Fagents --config-output .\u002Frouting-test.json\n```\n\n**Shared memory corruption?**\n```bash\npython -m json.tool shared_memory\u002Fshared_data.json  # Validate JSON\npython scripts\u002Fmemory_sync.py --cleanup              # Remove expired\npython scripts\u002Fmemory_sync.py --list-schemas         # See available typed schemas\n```\n\n**Fleet health check?**\n```bash\npython scripts\u002Ffleet_validate.py --agents .\u002Fagents --config ~\u002F.openclaw\u002Fconfig.json\n```\n\n**Schema validation on write?**\n```bash\npython scripts\u002Fmemory_sync.py --write 'coordinator:task:{\"task_id\":\"t1\",\"assigned_to\":\"research\"}' --schema task_assignment\n```\n\n---\n\n## Version History\n\n| Version | Date | Changes |\n|---------|------|--------|\n| 1.4.0 | 2026-05-27 | **HTTP REST API** — `--listen` flag starts zero-dep REST server exposing shared memory over HTTP. GET\u002FPOST for read, write, keys, stats, count, cleanup, schemas, health. CORS headers, stdlib only |\n| 1.3.0 | 2026-05-26 | **SQLite backend** — WAL-mode concurrent shared memory, ACID transactions, `--backend sqlite`, `--migrate-to-sqlite` tool, `--keys` and `--count-by-agent` utilities, `--compact` for space reclamation. No pip deps required. SKILL.md and memory-schemas.md updated with SQLite docs |\n| 1.2.0 | 2026-05-25 | Structured memory schemas with validation; health probes in routing config; memory-schemas.md reference; rate limiting; `--schema` and `--list-schemas` flags; **workflows.md** with 5 DAG templates; schema validation bug fix (inner data fields + correct schema type); reference list expanded |\n| 1.1.0 | 2026-05-23 | Fleet validation script added; deployment config templates shipped; date bug fixed; OpenClaw 2026.3+ pattern notes |\n| 1.0.1 | 2026-05-11 | Templates unified; agent_setup.py reads from assets\u002Ftemplates\u002F; _meta.json + origin tracking added |\n| 1.0.0 | 2026-04-02 | Initial release — agent setup, routing config, memory sync, cloud deploy |\n\n---\n\n*Based on production patterns from Abhi's multi-agent deployment experience. Tested with OpenClaw 2026.2+.*\n*Published on [ClawHub](https:\u002F\u002Fclawhub.ai) — slug: multi-agent-deployment*",{"name":5,"description":7},"master","4aa004da461d820f4ea5499e58eb8d196cf99016","community\u002Fclawhub\u002Fs\u002Fskill-multi-agent-deployment",0,"https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge","https:\u002F\u002Fskill-explorer-rho.vercel.app","https:\u002F\u002Fgithub.com\u002FLord1Egypt","Unbreakable. Unstoppable. Unmatched⚡.\r\n Speaker of machine tongues 🤖  | World-builder of decentralized dreams 🌐 | Defender of the digital realm🛡️.","2026-06-10 08:21:01",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add Lord1Egypt\u002Fawesome-skill-forge --skill multi-agent-deployment",{"tool":39,"command":40},"bunx","bunx skills add Lord1Egypt\u002Fawesome-skill-forge --skill multi-agent-deployment",{"tool":42,"command":43},"pnpm","pnpm dlx skills add Lord1Egypt\u002Fawesome-skill-forge --skill multi-agent-deployment","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fcommunity\u002Fclawhub\u002Fs\u002Fskill-multi-agent-deployment\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M785WKJB99G3STR1NF9NRTJW","youtube-content","YouTube transcripts to summaries, threads, blogs.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fyoutube-content\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M785WEFHB0G2S9YA6EPR3H0E","xurl","X\u002FTwitter via xurl CLI: post, search, DM, media, v2 API.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fxurl\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M785W98EDN4EFJXJPMQEHCRC","touchdesigner-mcp","Control a running TouchDesigner instance via twozero MCP — create operators, set parameters, wire connections, execute Python, build real-time visuals. 36 native tools.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftouchdesigner-mcp\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M785W5S9FP3QBH2V191E025P","test-driven-development","TDD: enforce RED-GREEN-REFACTOR, tests before code.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Ftest-driven-development\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M785W2AS08NSSH9YATCV6FJB","teams-meeting-pipeline","Operate the Teams meeting summary pipeline via Hermes CLI — summarize meetings, inspect pipeline status, replay jobs, manage Microsoft Graph subscriptions.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fteams-meeting-pipeline\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M785VXKW4T34BTP1SCHW1R7F","systematic-debugging","4-phase root cause debugging: understand bugs before fixing.","https:\u002F\u002Fgithub.com\u002FLord1Egypt\u002Fawesome-skill-forge\u002Fblob\u002Fmaster\u002Fskills\u002Fsystematic-debugging\u002FSKILL.md",[],[]]