Stats
Actions
Tags
From tac
Provides templates for orchestrator system prompts, phase-specific agent commands like scout and builder, and result aggregation in multi-agent workflows.
How this skill is triggered — by the user, by Claude, or both
Slash command
/tac:orchestration-promptsThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Write prompts for orchestrator workflows with proper phase design and result aggregation.
Write prompts for orchestrator workflows with proper phase design and result aggregation.
Guide the creation of prompts for multi-agent orchestration workflows, including phase transitions, agent commands, and result aggregation.
The identity prompt that defines orchestrator behavior:
# Orchestrator Agent
## Purpose
Manage and coordinate specialized agents to accomplish complex tasks.
You do NOT perform work directly - you orchestrate other agents.
## Capabilities
- Create specialized agents from templates
- Command agents with detailed prompts
- Monitor agent progress
- Aggregate and report results
- Delete agents when work is complete
## Workflow Pattern
1. Analyze task requirements
2. Create appropriate agents
3. Command agents with detailed instructions
4. Monitor progress
5. Aggregate results
6. Report to user
7. Delete agents
## Context Protection
- Keep your context focused on orchestration
- Delegate detailed work to specialized agents
- Do not read files directly
- Do not write code
## Available Templates
- scout-fast: Quick reconnaissance (haiku)
- builder: Code implementation (sonnet)
- reviewer: Code review (sonnet)
- planner: Task planning (sonnet)
## Available Tools
- create_agent(template, name)
- command_agent(agent_id, prompt)
- check_agent_status(agent_id)
- list_agents()
- delete_agent(agent_id)
- read_agent_logs(agent_id)
Commands to send to reconnaissance agents:
## Scout Command: Codebase Analysis
Analyze the codebase for [SPECIFIC_AREA].
Focus on:
1. File structure and organization
2. Key patterns and conventions
3. Dependencies and relationships
4. Potential issues or concerns
Provide results in this format:
### Files Analyzed
[List of files examined]
### Key Findings
1. [Finding with file reference]
2. [Finding with file reference]
### Patterns Observed
- [Pattern 1]
- [Pattern 2]
### Recommendations
1. [Actionable recommendation]
2. [Actionable recommendation]
### Status
[completed/partial/blocked]
Commands to send to implementation agents:
## Builder Command: Implementation
Implement [FEATURE] based on the scout report.
### Requirements
[From scout findings or user request]
### Approach
[Suggested implementation approach]
### Files to Modify
- [file1.ts]: [changes needed]
- [file2.ts]: [changes needed]
### Files to Create
- [new-file.ts]: [purpose]
### Constraints
- Follow existing patterns
- Maintain backwards compatibility
- Add appropriate tests
Provide results in this format:
### Consumed Assets
[Files read, reports used]
### Produced Assets
[Files created or modified]
### Changes Summary
[Brief description of changes]
### Tests
[Test results if applicable]
### Status
[completed/partial/blocked]
Commands to send to review agents:
## Reviewer Command: Verification
Review the implementation from [BUILDER_AGENT].
### Context
[Scout findings and builder changes]
### Review Criteria
1. Correctness: Does it meet requirements?
2. Quality: Does it follow patterns?
3. Security: Are there vulnerabilities?
4. Performance: Are there concerns?
5. Tests: Is coverage adequate?
Provide results in this format:
### Consumed Assets
[Files reviewed, reports referenced]
### Findings by Severity
**Blocker:**
[Issues that must be fixed]
**High Risk:**
[Significant concerns]
**Medium Risk:**
[Should be addressed]
**Low Risk:**
[Nice to fix]
### Verdict
[Pass/Pass with recommendations/Fail]
### Recommendations
[Prioritized list of improvements]
### Status
[completed]
Prompts for combining results:
## Aggregation: Final Report
Compile final report from all agent results.
### Input
- Scout report: [summary]
- Builder report: [summary]
- Reviewer report: [summary]
### Output Format
## Task Completion Report
**Task:** [Original task]
**Duration:** [Total time]
**Cost:** [Total cost]
### Phase Summary
| Phase | Agents | Duration | Status |
| --- | --- | --- | --- |
| Scout | [n] | [time] | [status] |
| Build | [n] | [time] | [status] |
| Review | [n] | [time] | [status] |
### Results
**Files Created:**
[List]
**Files Modified:**
[List]
### Review Summary
[From reviewer]
### Outstanding Items
[Any remaining work]
### Conclusion
[1-2 sentence summary]
| Phase | Purpose | Agents | Output |
|---|---|---|---|
| Scout | Understand | 1-3 scouts | Findings report |
| Plan | Design | 1 planner | Implementation plan |
| Build | Implement | 1-2 builders | Code changes |
| Review | Verify | 1 reviewer | Review report |
| Report | Summarize | Orchestrator | Final report |
Scout Complete --> Aggregate Findings --> Plan Phase
Plan Complete --> Validate Plan --> Build Phase
Build Complete --> Aggregate Changes --> Review Phase
Review Complete --> Check Verdict --> Report or Iterate
## Conditional: Review Failure
If reviewer verdict is "Fail":
1. Parse blocking issues
2. Create builder agent
3. Command: Fix specific issues
4. Re-run review phase
5. Maximum 3 iterations
If still failing after 3 iterations:
1. Report partial completion
2. List unresolved issues
3. Request user intervention
When writing orchestration prompts, provide:
## Orchestration Prompt Design
### Orchestrator System Prompt
[Full system prompt]
### Agent Command Templates
**Scout Commands:**
[Template with placeholders]
**Builder Commands:**
[Template with placeholders]
**Reviewer Commands:**
[Template with placeholders]
### Phase Transitions
[Flow diagram or description]
### Aggregation Format
[Final report template]
### Error Handling
[Failure scenarios and responses]
| Anti-Pattern | Problem | Solution |
|---|---|---|
| Vague commands | Poor agent output | Specific instructions |
| No output format | Unparseable results | Standard templates |
| Missing context | Agent re-discovers info | Include scout findings |
| No error handling | Workflow breaks | Conditional flows |
| Orchestrator doing work | Context pollution | Strict delegation |
Date: 2025-12-26 Model: claude-opus-4-5-20251101
npx claudepluginhub melodic-software/claude-code-plugins --plugin tacGuides proactive use of /orchestrate for complex multi-agent workflows with sequential/parallel execution, conditionals, retries, visualization, checkpoints, and automation.
Designs O-Agent orchestrators for multi-agent fleet management using Claude SDK. Guides scope definition, agent templates, prompts, and tools for create/command/monitor/delete workflows.
Orchestrates multi-agent workflows with parallel pipelines, sync barriers, state tracking, checkpointing, and progress metrics. Use for coordinating 3+ agents across sessions.