{"id":38780427,"url":"https://github.com/xuzeyu91/distributed-task-orchestrator","last_synced_at":"2026-01-18T17:00:46.393Z","repository":{"id":332082122,"uuid":"1132663664","full_name":"xuzeyu91/distributed-task-orchestrator","owner":"xuzeyu91","description":"A powerful agent skill for orchestrating complex, multi-step tasks through distributed sub-agent execution. This skill decomposes complex user requests into atomic tasks and manages parallel execution through virtual agents, with optional integration with Claude CLI for true distributed processing.","archived":false,"fork":false,"pushed_at":"2026-01-12T14:19:03.000Z","size":47,"stargazers_count":3,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2026-01-12T18:52:34.228Z","etag":null,"topics":["agents","claude-code","skills"],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/xuzeyu91.png","metadata":{"files":{"readme":"readme.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-01-12T09:27:01.000Z","updated_at":"2026-01-12T14:19:07.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/xuzeyu91/distributed-task-orchestrator","commit_stats":null,"previous_names":["xuzeyu91/distributed-task-orchestrator"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/xuzeyu91/distributed-task-orchestrator","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/xuzeyu91%2Fdistributed-task-orchestrator","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/xuzeyu91%2Fdistributed-task-orchestrator/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/xuzeyu91%2Fdistributed-task-orchestrator/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/xuzeyu91%2Fdistributed-task-orchestrator/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/xuzeyu91","download_url":"https://codeload.github.com/xuzeyu91/distributed-task-orchestrator/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/xuzeyu91%2Fdistributed-task-orchestrator/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":28543495,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-01-18T14:59:57.589Z","status":"ssl_error","status_checked_at":"2026-01-18T14:59:46.540Z","response_time":98,"last_error":"SSL_read: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agents","claude-code","skills"],"created_at":"2026-01-17T12:24:01.495Z","updated_at":"2026-01-18T17:00:46.380Z","avatar_url":"https://github.com/xuzeyu91.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"# Distributed Task Orchestrator\n\nA powerful agent skill for orchestrating complex, multi-step tasks through distributed sub-agent execution. This skill decomposes complex user requests into atomic tasks and manages parallel execution through virtual agents, with optional integration with Claude CLI for true distributed processing.\n\n## Table of Contents\n\n- [Overview](#overview)\n- [When to Use This Skill](#when-to-use-this-skill)\n- [Quick Start](#quick-start)\n- [Core Workflow](#core-workflow)\n- [File Structure](#file-structure)\n- [Usage Scenarios](#usage-scenarios)\n- [Claude CLI Integration](#claude-cli-integration)\n- [Templates](#templates)\n- [Best Practices](#best-practices)\n- [Related Documentation](#related-documentation)\n\n---\n\n## Overview\n\nThe Distributed Task Orchestrator transforms how you handle complex, multi-step tasks by:\n\n1. **Decomposing** complex requests into independent atomic tasks\n2. **Assigning** virtual agents to handle each task\n3. **Executing** tasks in parallel (simulated or via CLI)\n4. **Aggregating** results into a cohesive final output\n\n### Key Features\n\n- **Task Decomposition**: Automatically breaks down complex requests into manageable atomic tasks\n- **Dependency Management**: Handles serial, parallel, and DAG dependencies between tasks\n- **Virtual Agent System**: Assigns Agent-01, Agent-02, etc. to execute each task\n- **Parallel Execution**: Maximizes efficiency by running independent tasks simultaneously\n- **State Persistence**: Uses markdown files to track progress and maintain context\n- **Claude CLI Integration**: Optionally launch real sub-agents via Claude CLI for true distributed processing\n- **Error Recovery**: Automatic retry and graceful failure handling\n\n---\n\n## When to Use This Skill\n\n### Recommended Use Cases\n\n\n| Trigger                      | Description                                           | Example                                               |\n| ---------------------------- | ----------------------------------------------------- | ----------------------------------------------------- |\n| **Complex multi-step tasks** | Tasks requiring 3+ distinct steps                     | \"Build a complete authentication system\"              |\n| **Parallel execution needs** | Independent tasks that can run simultaneously         | \"Translate 5 documents to Chinese\"                    |\n| **Sub-agent orchestration**  | Need to coordinate multiple specialized agents        | \"Analyze code for quality, security, and performance\" |\n| **Large-scale analysis**     | Tasks requiring analysis of multiple files/components | \"Review all API endpoints in the project\"             |\n| **Task decomposition**       | Complex requests needing systematic breakdown         | \"Refactor the entire frontend architecture\"           |\n\n\n### Trigger Keywords\n\nThe skill activates when users mention:\n\n- \"parallel\", \"simultaneously\", \"concurrent\"\n- \"subtask\", \"sub-task\", \"atomic task\"\n- \"agent\", \"sub-agent\", \"Agent-01\"\n- \"orchestrate\", \"coordinate\", \"distribute\"\n- \"break down\", \"decompose\"\n\n### When NOT to Use\n\n- Simple, single-step tasks\n- Tasks completable in \u003c 3 steps\n- Tasks with strict sequential dependencies only\n- Quick questions or clarifications\n\n---\n\n## Quick Start\n\n### 1. Initialize the Orchestration Directory\n\n```bash\nmkdir .orchestrator\nmkdir .orchestrator/agent_tasks\nmkdir .orchestrator/results\n```\n\n### 2. Create the Master Plan\n\nCreate `.orchestrator/master_plan.md` with your task decomposition:\n\n```markdown\n# Task Plan\n\n## Original Request\n\u003e Analyze my TypeScript project for code quality, security, and performance\n\n## Task Decomposition\n| Task ID | Description | Dependencies | Agent |\n|---------|-------------|--------------|-------|\n| T-01 | Scan source files | None | Agent-01 |\n| T-02 | Check code quality | T-01 | Agent-02 |\n| T-03 | Security scan | T-01 | Agent-03 |\n| T-04 | Performance analysis | T-01 | Agent-04 |\n| T-05 | Generate report | T-02,T-03,T-04 | Agent-05 |\n```\n\n### 3. Execute and Monitor\n\nThe orchestrator will:\n\n1. Execute T-01 first (no dependencies)\n2. Run T-02, T-03, T-04 in parallel (all depend only on T-01)\n3. Execute T-05 after all parallel tasks complete\n4. Generate final output in `.orchestrator/final_output.md`\n\n---\n\n## Core Workflow\n\n### Phase 1: Task Analysis and Decomposition\n\n1. **Analyze user intent** - Identify primary and secondary goals\n2. **Build dependency graph** - Map relationships between tasks\n3. **Create atomic tasks** - Each task should be independently executable\n4. **Define I/O specifications** - Clear inputs and expected outputs\n\n### Phase 2: Agent Assignment and Status Marking\n\n1. **Assign Agent IDs** - Agent-01, Agent-02, etc.\n2. **Create task status table** - Track progress\n3. **Generate task files** - `.orchestrator/agent_tasks/agent-XX.md`\n4. **Initialize status** - 🟡 Pending\n\n**Status Icons:**\n\n- 🟡 Pending - Awaiting execution\n- 🔵 Running - Currently executing\n- ✅ Completed - Execution successful\n- ❌ Failed - Execution failed\n- ⏸️ Waiting - Dependencies not satisfied\n- 🔄 Retrying - Retry in progress\n\n### Phase 3: Parallel Execution\n\n**Method A: Simulated Execution**\n\n```\n═══════════════════════════════════════════════════\n🤖 Agent-01 [T-01: Read Code]\n───────────────────────────────────────────────────\n📥 Instruction: Read all TypeScript files\n⚙️ Execution:\n   1. Scan directory structure\n   2. Read file contents\n📤 Output: File statistics saved\n✅ Status: Completed\n═══════════════════════════════════════════════════\n```\n\n**Method B: Claude CLI Execution**\n\n```powershell\n# Launch sub-agents via Claude CLI\n$task = Get-Content \".orchestrator/agent_tasks/agent-01.md\" -Raw\nclaude -p $task | Out-File \".orchestrator/results/agent-01-result.md\"\n```\n\n### Phase 4: Result Aggregation\n\n1. **Collect all results** - From `.orchestrator/results/`\n2. **Verify completeness** - Ensure all tasks completed\n3. **Merge by dependency order** - Combine results logically\n4. **Generate final output** - `.orchestrator/final_output.md`\n\n---\n\n## File Structure\n\n### Skill Files (Reference)\n\n```\ndistributed-task-orchestrator/\n├── SKILL.md              # Main skill entry point\n├── workflow.md           # Detailed workflow documentation\n├── templates.md          # Task and status templates\n├── cli-integration.md    # Claude CLI integration guide\n├── examples.md           # Practical examples\n├── notes.md              # Design notes\n└── task_plan.md          # Development task plan\n```\n\n### Runtime Files (Generated in User's Project)\n\n```\n[user-project]/\n├── .orchestrator/\n│   ├── master_plan.md          # Master task plan and status\n│   ├── agent_tasks/\n│   │   ├── agent-01.md         # Agent-01 task description\n│   │   ├── agent-02.md         # Agent-02 task description\n│   │   └── ...\n│   ├── results/\n│   │   ├── agent-01-result.md  # Agent-01 execution result\n│   │   ├── agent-02-result.md  # Agent-02 execution result\n│   │   └── ...\n│   └── final_output.md         # Aggregated final output\n```\n\n---\n\n## Usage Scenarios\n\n### Scenario 1: Codebase Analysis\n\n**Request:** \"Analyze my TypeScript project for code quality, security, and performance\"\n\n**Task Decomposition:**\n\n```\n[T-01: Code Scan] ──┬──→ [T-02: Quality Analysis]\n                    ├──→ [T-03: Security Scan]\n                    └──→ [T-04: Performance Analysis]\n                                   ↓\n                    [T-05: Generate Report] ←─────┘\n```\n\n**Execution Flow:**\n\n1. T-01 reads all source files (1.8s)\n2. T-02, T-03, T-04 run in parallel (3.2s max)\n3. T-05 aggregates all findings (1.5s)\n4. Total: ~6.5s (vs ~11s sequential)\n\n### Scenario 2: Multi-Document Translation\n\n**Request:** \"Translate the 5 documents in docs/ to Chinese\"\n\n**Task Decomposition:**\n\n```\n[T-01: intro.md] ─────────────────────→ All independent\n[T-02: getting-started.md] ───────────→ Maximum parallelism\n[T-03: api-reference.md] ─────────────→ \n[T-04: tutorials.md] ─────────────────→\n[T-05: faq.md] ───────────────────────→\n```\n\n**Execution:**\n\n- All 5 tasks run simultaneously\n- Total time: ~45s (vs ~180s sequential)\n- Parallel efficiency: 4x speedup\n\n### Scenario 3: API Endpoint Testing\n\n**Request:** \"Test all API endpoints for response time and correctness\"\n\n**Task Decomposition:**\n\n- T-01: Test GET /api/users\n- T-02: Test GET /api/users/:id\n- T-03: Test POST /api/users\n- T-04: Test GET /api/products\n- T-05: Test GET /api/orders\n\n**Final Report Includes:**\n\n- Endpoints tested: 5\n- Test cases: 15\n- Pass rate: 93%\n- Average response time: 156ms\n- Failed cases with recommendations\n\n### Scenario 4: Full-Stack Feature Development\n\n**Request:** \"Add user authentication with JWT\"\n\n**Task Decomposition:**\n\n```\n[T-01: Design DB Schema] ───→ [T-03: Implement User Model]\n[T-02: Design API Spec] ────→ [T-04: Implement Auth Routes]\n                                        ↓\n                            [T-05: Create JWT Middleware]\n                                        ↓\n                            [T-06: Add Frontend Auth]\n                                        ↓\n                            [T-07: Write Tests]\n```\n\n---\n\n## Claude CLI Integration\n\n### Basic Commands\n\n```powershell\n# Direct execution\nclaude -p \"Your task description\"\n\n# Read task from file\nclaude -p (Get-Content task.md -Raw)\n\n# Save result to file\nclaude -p \"Task description\" | Out-File result.md\n\n# JSON output format\nclaude -p \"Task description\" --output-format json\n```\n\n### Parallel Execution with PowerShell Jobs\n\n```powershell\n# Launch multiple agents in parallel\n$taskFiles = Get-ChildItem \".orchestrator/agent_tasks/*.md\"\n\n$jobs = foreach ($file in $taskFiles) {\n    $agentId = $file.BaseName\n    Start-Job -Name $agentId -ScriptBlock {\n        param($taskPath, $resultPath)\n        $task = Get-Content $taskPath -Raw\n        claude -p $task | Out-File $resultPath -Encoding UTF8\n    } -ArgumentList $file.FullName, \".orchestrator/results/$agentId-result.md\"\n}\n\n# Wait for all to complete\n$jobs | Wait-Job\n\n# Display results\n$jobs | ForEach-Object {\n    $status = if ($_.State -eq 'Completed') { \"✅\" } else { \"❌\" }\n    Write-Host \"$status $($_.Name)\"\n}\n\n# Cleanup\n$jobs | Remove-Job\n```\n\n### Dependency-Aware Execution\n\n```powershell\n# Define task dependencies\n$taskGraph = @{\n    \"T-01\" = @{ Agent = \"Agent-01\"; Deps = @() }\n    \"T-02\" = @{ Agent = \"Agent-02\"; Deps = @(\"T-01\") }\n    \"T-03\" = @{ Agent = \"Agent-03\"; Deps = @(\"T-01\") }\n    \"T-04\" = @{ Agent = \"Agent-04\"; Deps = @(\"T-02\", \"T-03\") }\n}\n\n# Execute in topological order with parallel batches\n```\n\n### Error Handling with Retry\n\n```powershell\nfunction Invoke-AgentWithRetry {\n    param(\n        [string]$TaskFile,\n        [string]$ResultFile,\n        [int]$TimeoutSeconds = 300,\n        [int]$MaxRetries = 3\n    )\n    \n    $retryCount = 0\n    while ($retryCount -lt $MaxRetries) {\n        $job = Start-Job -ScriptBlock {\n            param($taskPath)\n            claude -p (Get-Content $taskPath -Raw)\n        } -ArgumentList $TaskFile\n        \n        if (Wait-Job $job -Timeout $TimeoutSeconds) {\n            $result = Receive-Job $job\n            Remove-Job $job\n            $result | Out-File $ResultFile -Encoding UTF8\n            return @{ Success = $true; Retries = $retryCount }\n        }\n        \n        Stop-Job $job\n        Remove-Job $job\n        $retryCount++\n    }\n    \n    return @{ Success = $false; Retries = $retryCount }\n}\n```\n\n---\n\n## Templates\n\n### Master Plan Template\n\n```markdown\n# 🎯 Distributed Task Plan\n\n## Original Request\n\u003e [User's original request]\n\n## Goal Definition\n**Primary Goal**: [Final result to achieve]\n**Success Criteria**: [How to determine completion]\n\n## Task Decomposition\n| Task ID | Task Name | Description | Dependencies | Priority |\n|---------|-----------|-------------|--------------|----------|\n| T-01 | [Name] | [Description] | None | P0 |\n\n## Agent Assignment\n| Task ID | Agent | Status | Start Time | End Time |\n|---------|-------|--------|------------|----------|\n| T-01 | Agent-01 | 🟡 Pending | - | - |\n```\n\n### Agent Task Template\n\n```markdown\n# 🤖 Agent-XX Task Assignment\n\n## Task Information\n- **Task ID**: T-XX\n- **Task Name**: [Name]\n- **Priority**: P1\n- **Estimated Time**: 3 minutes\n\n## Input\n| Parameter | Type | Source | Value |\n|-----------|------|--------|-------|\n| param1 | string | User input | [Value] |\n\n## Task Description\n[Detailed task description]\n\n## Expected Output\n[Format and content requirements]\n\n## Constraints\n- [Constraint 1]\n- [Constraint 2]\n```\n\n### Agent Result Template\n\n```markdown\n# 📤 Agent-XX Execution Result\n\n## Execution Summary\n- **Status**: ✅ Success\n- **Duration**: X.Xs\n\n## Output\n[Actual output content]\n\n## Statistics\n| Metric | Value |\n|--------|-------|\n| Items processed | X |\n| Successful | X |\n```\n\n---\n\n## Best Practices\n\n### 1. Task Granularity\n\n- **Ideal**: Each task completes in 1-5 minutes\n- **Too large**: Break down further\n- **Too small**: Consider merging\n\n### 2. Minimize Dependencies\n\n- Design independent tasks whenever possible\n- Use files to pass intermediate results\n- Avoid circular dependencies\n\n### 3. Maximize Parallelism\n\n- More independent tasks = greater speedup\n- Recommended concurrency: 4-8 agents\n- Monitor resource usage\n\n### 4. State Persistence\n\n- Update `master_plan.md` after every state change\n- Use file system as external memory\n- Log all executions and errors\n\n### 5. Error Handling\n\n- Isolate failures - one agent's failure shouldn't block others\n- Implement automatic retry with exponential backoff\n- Preserve partial results for recovery\n\n### 6. Dependency Types\n\n\n| Type            | Description            | Example           |\n| --------------- | ---------------------- | ----------------- |\n| Data Dependency | B needs A's output     | Analyze → Report  |\n| Sequential      | B must follow A        | Create → Populate |\n| Resource        | A and B share resource | Same file writes  |\n| None            | Completely independent | Different files   |\n\n\n---\n\n## Related Documentation\n\n- [workflow.md](distributed-task-orchestrator/workflow.md) - Detailed workflow documentation\n- [templates.md](distributed-task-orchestrator/templates.md) - Complete template collection\n- [cli-integration.md](distributed-task-orchestrator/cli-integration.md) - Claude CLI deep integration\n- [examples.md](distributed-task-orchestrator/examples.md) - Practical examples\n- [notes.md](distributed-task-orchestrator/notes.md) - Design notes\n\n---\n\n## Installation\n\nThis skill is located at:\n\n```\nC:\\Users\\{User}\\.claude\\skills\\distributed-task-orchestrator\\SKILL.md\n```\n\nThe skill is automatically available when:\n\n- User needs to orchestrate complex multi-step tasks\n- User mentions parallel execution or sub-agents\n- User needs to launch Claude CLI for subtasks\n- User requests task decomposition\n\n---\n\n## Contributing\n\nTo extend or modify this skill:\n\n1. Edit files in the `distributed-task-orchestrator/` directory\n2. Update `SKILL.md` for core workflow changes\n3. Add new templates to `templates.md`\n4. Document new CLI patterns in `cli-integration.md`\n5. Add examples to `examples.md`\n\n---\n\n## License\n\nThis skill is part of a personal knowledge base and skill collection.","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fxuzeyu91%2Fdistributed-task-orchestrator","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fxuzeyu91%2Fdistributed-task-orchestrator","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fxuzeyu91%2Fdistributed-task-orchestrator/lists"}