# Coordinate Parallel AI Subagents

Complex projects often contain independent tasks that waste time when handled sequentially. This skill shows how to plan, launch, collect, and synthesize parallel subagent work.

## Install

```bash
npx skillstore add cloudai-x/parallel-execution
```

## Metadata

- Status: approved
- Slug: cloudai-x-parallel-execution
- Skillstore revision: r1
- Version status: missing
- Tree hash: 2e1136693e5708719ed74c8d1498a4824ca7efc0ce76738f22636949524a0792
- Author: CloudAI-X
- GitHub username: CloudAI-X
- License: MIT
- Repository: https://github.com/CloudAI-X/claude-workflow/tree/main/skills/parallel-execution
- Ref: d11af420e07cd866fe39cfa937efef3223896c06
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: external\_commands
- Quality score: 78
- Quality tier: bronze
- Public page: https://skillstore.pages.dev/skills/cloudai-x-parallel-execution
- Manifest: https://skillstore.pages.dev/api/skills/cloudai-x-parallel-execution/manifest

## Capabilities

- Explains when independent tasks can safely run at the same time.
- Shows how to launch multiple Task calls in one assistant message.
- Describes result retrieval with TaskOutput and synthesis steps.
- Provides patterns for task, directory, and perspective based parallelization.
- Lists cases where sequential work is safer than parallel work.

## Use Cases

- Parallel Feature Delivery: Split independent implementation, testing, and documentation tasks across specialized subagents, then combine the results.
- Multi Angle Code Review: Run security, performance, testing, and architecture reviews at the same time for faster feedback.
- Repository Exploration: Assign different directories or modules to separate subagents and synthesize their findings into one overview.

## Prompt Templates

### Identify Parallel Tasks

```
Review this request and list which tasks can run in parallel. Mark any task that must wait for another result.
```

### Prepare Subagent Prompts

```
Create one focused subagent prompt for each independent task. Include role, context, files, expected output, and focus areas.
```

### Coordinate Parallel Review

```
Design a parallel review plan with security, performance, testing, and architecture subagents. Explain how their results should be merged.
```

### Synthesize Conflicting Results

```
Compare these subagent results, resolve conflicts, remove duplicates, and produce one prioritized implementation plan with risks and next actions.
```

## Limitations

- Does not run subagents or manage task state by itself.
- Requires a tool environment that supports background Task calls.
- Cannot prevent merge conflicts when tasks edit shared files.
- Assumes the orchestrator reviews and reconciles all outputs.

## Best Practices

- Confirm tasks are independent before launching background subagents.
- Give each subagent narrow scope, relevant files, and a clear output format.
- Always synthesize results and resolve conflicts before presenting final work.

## Anti Patterns

- Do not parallelize tasks that edit the same files without coordination.
- Do not launch dependent tasks before prerequisite results are available.
- Do not forward raw subagent outputs without review and deduplication.

## Security Audit

- Audited at: 2026-07-06T07:39:45.718\+00:00
- Summary: All static findings are false positives from Markdown backticks, placeholder task examples, and TaskOutput sample text in SKILL.md. I found no evidence of command execution, system reconnaissance, data exfiltration, prompt injection, or unsafe host access. The skill is instructional content only.

## Stats

- Views: 239
- Downloads: 9
- Favorites: 0
- Popularity score: 0