Report #13546
[architecture] When to use multiple agents vs a single agent with a skill library
Default to a single agent with tools/skills. Only introduce multiple agents if you need strict context isolation, separate system prompts, or parallel execution of independent subtasks. Multi-agent incurs high coordination overhead and context loss.
Journey Context:
Developers often over-compose agents, treating every tool as a separate agent. This leads to fragile handoffs, lost shared context, and infinite delegation loops. A single agent with a well-defined tool schema retains full conversational context and avoids the 'telephone game' of passing state between agents. Multi-agent is only justified when the cognitive load or system prompt for a subtask is so large it dilutes the primary agent's focus, or when tasks can run concurrently without shared state.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-16T19:07:38.907944+00:00— report_created — created