liveBook · Manning

Tool confirmation decision

As explained in Build an AI Agent (From Scratch)

A tool confirmation decision is a user's approval or rejection, optionally with modified arguments, for a pending tool call.

A tool confirmation decision approves or rejects a pending tool call before execution. The decision can be selective across a batch, and approved calls may carry modified arguments. section 6.3.7

A confirmation-aware agent follows this execution flow:

  • Store approval-required calls in execution state without executing them or creating results. section 6.3.7

Selective approval

When an agent requests three tool calls, a user can approve two and reject one. File deletion and payment processing must wait for approval because executing first cannot be undone. section 6.3.7

How Build an AI Agent (From Scratch) covers it

How deep it goes
Explains state storage, pause-and-resume control flow, identifier-based matching, and confirmation processing without detailing the full tool implementation.
Examples
Uses file deletion, payment processing, and a three-call batch with two approvals and one rejection.
What you can do after
Build an agent loop that pauses for confirmations, resumes safely, and executes only approved pending calls.
How it connects
Connects tool calls with execution state, agent loops, pause and resume, and tool results.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build an AI Agent (From Scratch). The section links point to the parts of the book it draws on.