Tool Inventory

The declared set of tools an agent may call; inventory size is a capability versus reliability and context-budget tradeoff.

A tool inventory is the full set of tools an LLM agent can use. Each entry typically declares a name, parameters, and documentation so the planner can select and call tools via function calling.

Sizing the inventory is a real design decision. More tools increase capability, but make correct selection harder and consume context with tool descriptions — which itself caps how many tools can practically be declared. Published agents span tiny inventories (a handful of tools) to thousands of APIs; the right size is found by tool selection ablation, not by maximal coverage. Within a conversation, drop tools known to be irrelevant to the current phase so they do not distract the planner under conversational agent context.

Without external tools, an agent is limited to what the base model can do alone (for example, generate text). With tools, actions split into perception (read-only) and environment change (write) under agent tool categories. Two practitioner positions pull the sizing decision in opposite directions: coding-agent builders argue for a minimal, general tool surface (a few tools plus a shell), while security guidance for agents that change live systems argues for a narrow action verb set with no raw shell; a common resolution is general tools for reading and narrow verbs for writing. Prefer clear, simple tool APIs: ambiguous or overly complex functions increase planning failure modes such as invalid parameters or wrong values.

---

Sources

• AI Engineering (Chip Huyen), ch. 6
• Prompt Engineering for LLMs (Berryman), ch. 8