Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
The dispatching-parallel-agents skill is designed to optimize the investigation of multiple independent failures within a system. When several test files or subsystems fail simultaneously, addressing them sequentially can lead to significant time wastage, particularly when each failure is unrelated and can be resolved independently. This skill enables the dispatch of separate agents for each issue, allowing them to be investigated concurrently, thus enhancing efficiency and expediting the resolution process.
Use this skill when you encounter three or more test files failing with different root causes, or when multiple subsystems are broken independently and can be addressed without shared state.
This skill is not suitable when failures are related, as fixing one may inadvertently resolve others, or when a full understanding of the system state is necessary.
The skill can be integrated with various testing frameworks as long as the tasks can be dispatched and run concurrently.
An effective agent prompt should be focused on a single problem domain, self-contained with all necessary context, and specific about the expected output.
Quick Setup:
.claude/skills/Repository
davila7/claude-code-templates