Should you have one flexible tool or six specific ones? This comes up on every agent project and there's a genuine trade-off rather than a right answer.

Fewer, Broader Tools	More, Narrower Tools
Fewer choices per step, so fewer wrong picks	Each one is easier to describe precisely
More flexible, handles cases you didn't foresee	Easier to set permissions on individually
Harder to describe well, so misuse is subtler	More tools crowds the choice at every step
Harder to restrict, one tool, many powers	Easy to audit what the agent can actually do
Good for read-only exploration	Good for anything that writes or costs money

The Rule That Works

Broad for reading, narrow for writing. A flexible search tool that can query several things is fine, the worst case is a wasted call. A flexible 'update_shipment' tool that can change any field is a permissions problem, because you can't approve part of it. Split write actions into specific, individually-controllable tools.

Agentic AI and MCP for Product Managers

Module 1What 'Agentic' Actually Means

Module 2How an Agent Actually Works

Module 3MCP: What It Is and Why You Should Care

Module 4Designing the Tool Layer

Module 5Permissions and the Human in the Loop

Module 6Evaluating Agents

Module 7Cost, Latency and Unpredictability

Module 8Designing What the User Sees

Module 9Risk and Blast Radius

Module 10When an Agent Is the Wrong Answer

Module 11Final Project: Design One Real Agent End to End

Module 12Appendix: The Prompt Library

Granularity