Dispatching Parallel Agents

When you've got multiple independent failures after a refactor, this is about spinning up separate agents to tackle each problem domain in parallel instead of debugging sequentially. Each agent gets a tightly scoped prompt with specific context, clear constraints, and an instruction to report back what they found. The key insight is recognizing when problems are truly independent (abort tests, batch completion, race conditions) versus related. In the documented example, three agents fixed six test failures concurrently across different files with zero conflicts. The pattern works when there's no shared state and you can clearly partition the work, but it requires discipline in crafting focused prompts and verifying the results don't step on each other.

Install to Claude Code

npx -y skills add ganyuanran/aegis --skill dispatching-parallel-agents --agent claude-code