How many agents can I run in parallel with this skill?

You can launch up to 15 agents per parallel batch. It is recommended to wait for one batch to complete before starting the next.

How do I monitor the progress of parallel tasks?

You can monitor progress by checking the status file in the .maestro/cache directory using commands like 'cat', 'wc -l', or 'tail -f' to see completion timestamps.

How does this skill prevent context bloat?

It avoids calling TaskOutput, which returns full agent results to the main context. Instead, it uses a background pattern where agents write status updates to local files.

Can agents share memory during parallel execution?

No, parallel agents are self-contained and do not share memory. All necessary context must be included in each specific task prompt.

Parallel Agent Orchestration

Name: Parallel Agent Orchestration
Author: scooter-lacroix

byscooter-lacroix

Productivity & Workflow

Optimizes multi-agent workflows by managing concurrent task execution while preventing context bloat.

About

This skill provides a robust architectural pattern for launching up to 15 concurrent subagents within the Claude Code environment, ensuring high performance without exhausting the context window. By utilizing background execution and a file-based status tracking system instead of heavy return values, it allows developers to scale complex research, testing, or backfill operations while maintaining real-time visibility through standardized logging and artifact management.

Key Features

Context-efficient background execution using the run_in_background pattern
Standardized output structures for research and exploration tasks
0 GitHub stars
Batch management protocols supporting up to 15 simultaneous agents
Lightweight file-based status monitoring to prevent token-heavy TaskOutput calls
Real-time progress tracking via append-only completion logs

Use Cases

Performing batch updates or backfills across different API providers in a single operation
Running parallel research or data exploration tasks across multiple sources simultaneously
Executing independent unit tests or code reviews across large codebases in parallel

About

Key Features

Context-efficient background execution using the run_in_background pattern
Standardized output structures for research and exploration tasks
0 GitHub stars
Batch management protocols supporting up to 15 simultaneous agents
Lightweight file-based status monitoring to prevent token-heavy TaskOutput calls
Real-time progress tracking via append-only completion logs

Use Cases

Performing batch updates or backfills across different API providers in a single operation
Running parallel research or data exploration tasks across multiple sources simultaneously
Executing independent unit tests or code reviews across large codebases in parallel