Your agent called the wrong tool.
Fix the description. Not the agent.
When an agent picks the wrong tool, the agent is rarely the problem — the tool registry is. Three rules that fix tool selection in production: name tools precisely, describe when each one should be used, and load only what the task needs.
Tags#ToolDesign#AIAgents#ToolSelection#MCP#PromptEngineering#Production
Key takeaways
- 1When an agent picks the wrong tool, the registry is broken — not the agent.
- 2Clear, specific tool names correct decisions; vague names produce wrong guesses.
- 3Every tool description needs a "when to use" trigger — that's what the model scores against.
- 4Load only the tools the task needs. Too many tools loaded = confusion + token waste.
- 5The tool is not the problem. Better descriptions create better decisions.
