Tool bundling
Tool bundling groups related tools behind a single high-level interface — so the agent calls one tool that does the right thing internally rather than choosing between many narrow tools.
When agents have too many narrow tools (get_user, get_user_email, get_user_phone, get_user_address) they mis-route. Tool bundling exposes a single high-level tool (get_user_info with an args schema specifying which fields to return) that handles the dispatch internally. Empirically reduces mis-routing rates and simplifies the agent's surface area without losing capability. Best practice in 2026 is to bundle tools by user-facing intent rather than by underlying API endpoint, even if it means writing more glue code on the server side.
When to use tool bundling
- Agents with many narrow tools that overlap in intent.
- Production agents where mis-routing is materially degrading quality.
Common mistakes
- Bundling too aggressively — the bundled tool's args schema becomes unwieldy.
- Not documenting the dispatch logic — debugging tool decisions becomes hard.
FAQ
What is tool bundling?
Tool bundling groups related tools behind a single high-level interface — so the agent calls one tool that does the right thing internally rather than choosing between many narrow tools.
When should I use tool bundling?
Agents with many narrow tools that overlap in intent. Production agents where mis-routing is materially degrading quality.
What are the most common mistakes with tool bundling?
Bundling too aggressively — the bundled tool's args schema becomes unwieldy. Not documenting the dispatch logic — debugging tool decisions becomes hard.
Related terms
- Tool use (LLM) — Tool use is the umbrella term for any LLM mechanism that lets the model invoke external functions, APIs, or services — function calling, code interpreter, MCP servers, browser actions.
- Tool router — A tool router is a layer in an agent that decides which tool to call (or which sub-agent to delegate to) for a given step — distinct from a model router which picks the underlying LLM.
- Tool shadowing — Tool shadowing is when two or more tools in an agent's toolkit overlap in purpose enough that the model routes ambiguously — usually picking the worse one or oscillating between them.
- Function calling (tool use) — Function calling lets a language model emit a structured request to invoke a developer-defined tool, enabling reliable JSON output and agent workflows.
Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/tool-bundling.md.