Agent skill

pipeline-orchestration

Master orchestrator for the entire development pipeline from Phase 1-12. Activates via codeword [ACTIVATE:PIPELINE_ORCHESTRATION_V1] injected by hooks when user wants to start automated development. Activation trigger: [ACTIVATE:PIPELINE_ORCHESTRATION_V1]

View SKILL.md on GitHub Repository
Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/pipeline-orchestration

SKILL.md

Pipeline Orchestration Skill

Activation Method

This skill activates when the hook system injects the codeword:

[ACTIVATE:PIPELINE_ORCHESTRATION_V1]

This occurs when:

  • User says "begin automated development" or "start pipeline"
  • Phase 4 is complete
  • User wants full automation from Phase 1-12

Pipeline Orchestrator Skill

What This Skill Does

The Pipeline Orchestrator is the master controller for fully automated development from Phase 1 through Phase 12. It:

  • Monitors completion signals from each phase
  • Automatically triggers the next phase skill
  • Handles errors gracefully with automatic recovery or human escalation
  • Preserves state through checkpointing
  • Provides progress dashboard with real-time status
  • Fully autonomous from Phase 6 through deployment (no manual gates)

When This Skill Activates

Primary Trigger: User completes Phase 0 (PRD creation, human validation, checkpoint passed)

Activation Phrases:

  • "Begin automated development"
  • "Start the pipeline"
  • "Automate phases 1 through 6"
  • "Run full development pipeline"
  • "Phase 4 is complete, start automation"

Prerequisites:

  • ✅ .taskmaster/scripts/phase4-checkpoint.sh passed
  • ✅ .taskmaster/tasks.json exists and validated
  • ✅ .taskmaster/docs/phase0-signoff.md exists
  • ✅ Git repository clean state

Pipeline Architecture

Phase 1-4 (Discovery & Planning) ────────────────────────┐
  ✅ Checkpoint Passed                   │ MANUAL
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
                ↓                        │
Phase 1: Task Decomposition             │ APPROVAL
  Skill: Task-Decomposer                │ (user approves tasks)
  Output: .signals/phase10-complete.json │
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
                ↓                        │
Phase 2: Spec Generation                │
  Skill: Spec-Generator                 │
  Output: .signals/phase11-complete.json │
                ↓                        │
Phase 3: Implementation                 │
  Skill: TDD-Implementer                │ FULLY
  Output: .signals/phase7-complete.json │ AUTONOMOUS
                ↓                        │
Phase 4: Integration Testing            │
  Skill: Integration-Tester             │
  Output: .signals/phase9-complete.json │
                ↓                        │
Phase 5: E2E & Production Validation    │
  Skill: E2E-Prod-Validator             │
  Output: .signals/phase10-complete.json │
                ↓                        │
Phase 6: Deployment & Infrastructure    │
  Skill: Deployment-Orchestrator        │
  Includes: Docker build & health check  │
  Output: .signals/phase11-complete.json │
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
                ↓
        ✅ DEPLOYED TO PRODUCTION

Completion Signal System

Each phase skill generates a completion signal file when done:

Phase 1 Signal

json
{
  "phase": 1,
  "phase_name": "Task Decomposition",
  "status": "success",
  "completed_at": "2025-10-29T15:30:00Z",
  "duration_minutes": 12,
  "summary": {
    "tasks_analyzed": 18,
    "tasks_expanded": 8,
    "subtasks_generated": 34,
    "high_complexity_tasks": [3, 5, 7, 9, 12, 15, 18, 22]
  },
  "next_phase": 2,
  "trigger_next": true
}

Phase 2 Signal

json
{
  "phase": 2,
  "phase_name": "Spec Generation",
  "status": "success",
  "completed_at": "2025-10-29T16:45:00Z",
  "duration_minutes": 45,
  "summary": {
    "tasks_processed": 18,
    "proposals_created": 15,
    "tightly_coupled": 10,
    "loosely_coupled": 5,
    "batches_completed": 3
  },
  "next_phase": 3,
  "trigger_next": true
}

Phase 3 Signal

json
{
  "phase": 3,
  "phase_name": "Implementation",
  "status": "success",
  "completed_at": "2025-10-29T19:30:00Z",
  "duration_minutes": 165,
  "summary": {
    "subtasks_implemented": 34,
    "tests_written": 187,
    "tests_passing": 187,
    "code_coverage": {
      "line": 87,
      "branch": 76
    },
    "implementation_strategy": "worktree-isolation"
  },
  "next_phase": 4,
  "trigger_next": true
}

Phase 4 Signal

json
{
  "phase": 4,
  "phase_name": "Integration Testing",
  "status": "success",
  "completed_at": "2025-10-29T20:15:00Z",
  "duration_minutes": 45,
  "summary": {
    "integration_points_total": 12,
    "integration_points_tested": 12,
    "integration_tests_passing": 12,
    "coverage": "100%"
  },
  "next_phase": 5,
  "trigger_next": true
}

Phase 5 Signal

json
{
  "phase": 5,
  "phase_name": "E2E & Production Validation",
  "status": "success",
  "completed_at": "2025-10-29T21:30:00Z",
  "duration_minutes": 75,
  "summary": {
    "e2e_workflows_tested": 8,
    "e2e_tests_passing": 8,
    "production_readiness_score": 94,
    "blocking_issues": 0
  },
  "next_phase": 6,
  "trigger_next": true
}

Phase 6 Signal

json
{
  "phase": 6,
  "phase_name": "Deployment",
  "status": "success",
  "completed_at": "2025-10-29T23:00:00Z",
  "duration_minutes": 90,
  "summary": {
    "staging_deployed": true,
    "staging_validated": true,
    "production_deployed": true,
    "deployment_strategy": "canary",
    "rollback_tested": true
  },
  "next_phase": null,
  "trigger_next": false,
  "pipeline_complete": true
}

Error Signal

json
{
  "phase": 3,
  "phase_name": "Implementation",
  "status": "error",
  "failed_at": "2025-10-29T18:15:00Z",
  "error": {
    "type": "test_failure",
    "message": "5 tests failing in user authentication module",
    "recovery_attempted": true,
    "recovery_successful": false,
    "requires_human": true
  },
  "checkpoint": ".taskmaster/.checkpoints/phase3-checkpoint-5.json",
  "next_action": "Fix failing tests, then resume from checkpoint"
}

Phase Transition Rules

Phase 1 → Phase 2 Transition

yaml
Triggers when:
  - ✅ .taskmaster/.signals/phase10-complete.json exists
  - ✅ status = "success"
  - ✅ trigger_next = true
  - ✅ All high-complexity tasks expanded

Action:
  - Load Phase 2 context
  - Activate Spec-Generator skill
  - Monitor for phase11-complete.json

Phase 2 → Phase 3 Transition

yaml
Triggers when:
  - ✅ .taskmaster/.signals/phase11-complete.json exists
  - ✅ status = "success"
  - ✅ trigger_next = true
  - ✅ All OpenSpec proposals created

Action:
  - Load Phase 3 context
  - Activate TDD-Implementer skill
  - Monitor for phase7-complete.json

Phase 3 → Phase 4 Transition

yaml
Triggers when:
  - ✅ .taskmaster/.signals/phase7-complete.json exists
  - ✅ status = "success"
  - ✅ trigger_next = true
  - ✅ All tests passing
  - ✅ Coverage ≥80% line, ≥70% branch

Action:
  - Load Phase 4 context
  - Activate Integration-Tester skill
  - Monitor for phase9-complete.json

Phase 4 → Phase 5 Transition

yaml
Triggers when:
  - ✅ .taskmaster/.signals/phase9-complete.json exists
  - ✅ status = "success"
  - ✅ trigger_next = true
  - ✅ 100% integration point coverage

Action:
  - Load Phase 5 context
  - Activate E2E-Prod-Validator skill
  - Monitor for phase10-complete.json

Phase 5 → Phase 6 Transition (AUTOMATIC)

yaml
Triggers when:
  - ✅ .taskmaster/.signals/phase10-complete.json exists
  - ✅ status = "success"
  - ✅ production_readiness_score ≥90%

Action:
  - Log Phase 5 summary
  - Automatically activate Deployment-Orchestrator skill
  - Pipeline proceeds to production deployment

Error Handling Strategy

Automatic Recovery (Skill Attempts First)

Category 1: Dependency Issues

yaml
Error: "npm: command not found"
Recovery:
  - Detect package manager needed
  - Install via appropriate method
  - Retry operation
  - If success: Continue
  - If failure: Escalate to human

Category 2: Test Failures

yaml
Error: "5 tests failing in authentication module"
Recovery:
  - Analyze test errors
  - Attempt automatic fixes (if safe):
    - Update test expectations
    - Fix obvious typos
    - Add missing mocks
  - If fixes succeed: Re-run tests
  - If still failing: Escalate to human with detailed report

Category 3: Transient Failures

yaml
Error: "Network timeout", "Rate limit hit"
Recovery:
  - Wait exponentially (1min, 2min, 4min)
  - Retry operation up to 3 times
  - If success: Continue
  - If still failing: Escalate to human

Human Escalation (Cannot Proceed)

Category 1: Ambiguous Decisions

yaml
Scenario: "Two equally valid implementation approaches"
Action:
  - Pause pipeline
  - Present options with pros/cons
  - Wait for human decision
  - Resume with selected approach

Category 2: Critical Errors

yaml
Scenario: "Git repository corrupted", "Database connection failed"
Action:
  - Save checkpoint immediately
  - Halt pipeline
  - Report error details
  - Provide recovery steps
  - Wait for human intervention

Category 3: Unknown Errors

yaml
Scenario: "Unexpected error not covered by recovery rules"
Action:
  - Save checkpoint
  - Log full error trace
  - Halt pipeline
  - Report to human with context

Checkpoint System

Checkpoint Files

.taskmaster/.checkpoints/
├── phase1-checkpoint-1.json  (after complexity analysis)
├── phase1-checkpoint-2.json  (after task 5 expanded)
├── phase1-checkpoint-3.json  (after task 10 expanded)
├── phase2-checkpoint-1.json  (after batch 1 complete)
├── phase2-checkpoint-2.json  (after batch 2 complete)
├── phase3-checkpoint-1.json  (after task 3 implemented)
├── phase3-checkpoint-2.json  (after task 7 implemented)
└── ...

Checkpoint Schema

json
{
  "phase": 2,
  "checkpoint_number": 3,
  "created_at": "2025-10-29T16:20:00Z",
  "state": {
    "current_operation": "Creating OpenSpec proposal for task #9",
    "completed_operations": [
      "Analyzed coupling for batch 1 (tasks 1-5)",
      "Created 5 OpenSpec proposals",
      "Analyzed coupling for batch 2 (tasks 6-10)",
      "Created 3 OpenSpec proposals"
    ],
    "pending_operations": [
      "Create proposal for task #9",
      "Create proposal for task #10",
      "Start batch 3 (tasks 11-15)"
    ]
  },
  "resume_command": "Continue Phase 2 from checkpoint 3: task #9"
}

Resume from Checkpoint

bash
# If pipeline interrupted, resume with:
# "Resume pipeline from last checkpoint"

# Orchestrator will:
1. Find most recent checkpoint
2. Load saved state
3. Resume from exact point of interruption
4. Continue pipeline execution

Progress Dashboard

While pipeline runs, orchestrator provides real-time updates:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
AUTOMATED DEVELOPMENT PIPELINE
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Phase 0: Setup & Planning            ✅ COMPLETE
Phase 1: Task Decomposition          ✅ COMPLETE (12 min)
Phase 2: Spec Generation             ✅ COMPLETE (45 min)
Phase 3: Implementation              🔄 IN PROGRESS (87 min)
  └─ Task 15/18 complete
  └─ 127/187 tests passing
  └─ Coverage: 82% line, 74% branch
Phase 4: Integration Testing         ⏸️  PENDING
Phase 5: E2E & Validation            ⏸️  PENDING
Phase 6: Deployment                  ⏸️  PENDING

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Elapsed Time: 144 minutes
Estimated Remaining: 156 minutes
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Usage Examples

Example 1: Full Pipeline Automation

User (after Phase 0 complete):

I've completed Phase 0 and the checkpoint passed. Begin automated development pipeline.

Orchestrator Response:

✅ Phase 4 Checkpoint Verified
   - tasks.json: 18 tasks validated
   - PRD: Complete
   - Sign-off: Obtained

🚀 Starting Automated Development Pipeline

Phase 1: Task Decomposition
  → Activating Task-Decomposer skill...
  → Analyzing complexity for 18 tasks...
  → Expanding high-complexity tasks...
  ✅ Phase 1 Complete (12 minutes)
     - 8 tasks expanded into 34 subtasks
     - Signal: phase10-complete.json generated

Phase 2: Spec Generation
  → Activating Spec-Generator skill...
  → Analyzing coupling for batch 1 (tasks 1-5)...
  → Creating OpenSpec proposals...
  → Batch 1 complete (5 proposals)
  → Processing batch 2...
  [continues automatically...]

Example 2: Error Recovery

Orchestrator detects error:

⚠️ Phase 3 Error Detected
   Error: 5 tests failing in user authentication module
   
🔄 Attempting Automatic Recovery
   → Analyzing test failures...
   → Identified: Missing mock for database connection
   → Adding mock configuration...
   → Re-running tests...
   ✅ Recovery Successful - All tests passing
   
Resuming Phase 3 Implementation...

Example 3: Human Escalation

Orchestrator encounters unknown error:

🚨 Phase 3 Error - Human Intervention Required

Error Details:
  Phase: 3 (Implementation)
  Operation: Implementing task #12
  Error: Unable to resolve dependency conflict
  
Checkpoint Saved: .taskmaster/.checkpoints/phase3-checkpoint-7.json

Options:
A) Manually resolve the dependency conflict
B) Skip task #12 for now (resume later)
C) Abort pipeline

Please respond: A, B, or C

Example 4: Phase 5 Auto-Transition

Orchestrator completes Phase 5:

✅ Phase 5 Complete: E2E & Production Validation

Production Readiness Score: 94%

Summary:
  - All E2E tests passing (8/8 workflows)
  - Integration tests: 100% coverage
  - Test coverage: 87% line, 76% branch
  - Security scans: PASSED
  - No blocking issues

🚀 Auto-proceeding to Phase 6: Deployment

[ACTIVATE:DEPLOYMENT_ORCHESTRATOR_V1]

Worktree Isolation Strategy

ALL development must happen in isolated git worktrees - NO exceptions.

Worktree Management

  • Each phase operates in its dedicated worktree: phase-X-task-Y
  • Orchestrator automatically creates worktrees before activating skills
  • Worktree boundaries are strictly enforced
  • No cross-worktree contamination allowed

Phase-Worktree Mapping

  1. Phase 1: Creates phase-10-task-1 worktree for task decomposition
  2. Phase 2: Creates individual worktrees per task batch: phase-11-task-N
  3. Phase 3: Each subtask gets isolated worktree: phase-7-task-N
  4. Phase 4: Integration testing in phase-9-task-1 worktree
  5. Phase 5: E2E validation in phase-10-task-1 worktree
  6. Phase 6: Deployment from phase-11-task-1 worktree

Worktree Lifecycle

bash
# Before each phase activation:
./lib/worktree-manager.sh create <phase> <task>
cd ./worktrees/phase-<phase>-task-<task>

# Phase execution with isolation enforcement
./hooks/worktree-enforcer.sh enforce

# After phase completion:
./lib/worktree-manager.sh merge phase-<phase>-task-<task>
./lib/worktree-manager.sh cleanup phase-<phase>-task-<task>

Skill Coordination

The orchestrator manages these phase skills with strict worktree isolation:

  1. Task-Decomposer (Phase 10) - Worktree: phase-10-task-1
  2. Spec-Generator (Phase 6) - Worktrees: phase-11-task-N per batch
  3. TDD-Implementer (Phase 7) - Worktrees: phase-7-task-N per subtask
  4. Integration-Tester (Phase 9) - Worktree: phase-9-task-1
  5. E2E-Prod-Validator (Phase 10) - Worktree: phase-10-task-1
  6. Deployment-Orchestrator (Phase 11) - Worktree: phase-11-task-1

Each skill:

  • Receives worktree context from orchestrator
  • Validates worktree isolation before execution
  • Executes its phase in dedicated worktree
  • Generates completion signal when done
  • Merges changes back to main branch
  • Cleans up worktree after successful merge

Success Criteria

Pipeline is successful when:

  • ✅ All 6 phases complete without errors
  • ✅ All tests passing (unit, integration, E2E)
  • ✅ Production readiness score ≥90%
  • ✅ GO decision from Phase 5
  • ✅ Successfully deployed to production
  • ✅ All validation gates passed

See Also

  • /DEVELOPMENT_WORKFLOW.md - Complete workflow documentation
  • /phase4-checkpoint.sh - Phase 0 verification script
  • /.taskmaster/.signals/ - Completion signal files
  • /.taskmaster/.checkpoints/ - Pipeline checkpoints

Didn't find tool you were looking for?

Be as detailed as possible for better results