ralph

Spawn isolated subagents per phase to orchestrate queue-based tasks.

1|Updated Feb 21, 2026
One-click install
npx skills add https://github.com/wtyler2505/ProtoPulse --skill ralph-wtyler2505
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ralph
Source: https://github.com/wtyler2505/ProtoPulse/tree/main/.agents/skills/ralph
Command: npx skills add https://github.com/wtyler2505/ProtoPulse --skill ralph-wtyler2505

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Queue-based task processing often risks context leakage across tasks when a single session handles multiple items. Ralph enforces isolation by spawning fresh subagents per phase to ensure clean context and robust handoffs.

Core Features & Use Cases

  • Per-task subagent spawning: each task runs in isolated context; prevents cross-task contamination.
  • Phase-driven control: supports serial, parallel, batch filtering, and dry-run modes.
  • Handoff-driven orchestration: requires RALPH HANDOFF blocks to communicate outcomes and queue updates.

Quick Start

Invoke /ralph N to start processing pending tasks with fresh subagent context.

Frequently Asked Questions about ralph

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I prevent context bleed when processing a queue of tasks?

Task orchestration with isolated subagents supports serial, parallel, and batch-filtered processing modes by enforcing phase progression and handoff parsing, ensuring correct queue updates and recoverable failures across each phase.

How do I start processing pending tasks with fresh subagent context?

To start processing pending tasks, invoke the /ralph command with the number of pending tasks to process, which triggers the spawning of isolated subagents to handle the queue with fresh context.

Can I run parallel and batch-filtered task processing without cross-task contamination?

Parallel and batch-filtered task processing runs without cross-task contamination by spawning isolated subagents for every subtask, using the Task tool to enforce strict context separation and prevent leakage across items.

What is the best way to manage handoffs between phases in task orchestration?

Managing handoffs between phases in task orchestration requires structured RALPH HANDOFF blocks to communicate outcomes and queue updates, allowing the next phase to resume with validated state and recoverable failures.

Does queue-based task orchestration support dry-run modes for phase management?

Queue-based task orchestration supports dry-run modes alongside serial, parallel, and batch filtering, allowing you to validate phase progression and handoff parsing without executing permanent queue updates.