the
∞
.ai
Search
/
Concepts
trail ·
0 nodes
Every concept
Tools
8 concepts, 8 reviewed and merged.
Evaluation
Agent Evaluation
Scoring an agent is harder than scoring an answer, because the same task has many valid trajectories and the thing…
verified
Interface
Computer Use
Instead of calling an API, drive the screen: look at a screenshot, move the mouse, type.
verified
Tool Schema
A tool is only usable by a model if the model can be told what it does, what it needs, and what it returns.
verified
Methods
Execution Feedback
The difference between an agent and a plan is that an agent gets to see what happened.
verified
Tool Selection
Given fifty tools and a request, which one — and is the answer none of them? Selection is a retrieval problem before…
verified
Regimes
Long-Horizon Agent
Tasks that take fifty steps fail differently from tasks that take five.
verified
Safety
Human in the Loop
Some actions should not happen without a person saying yes — sending money, deleting data, emailing a customer.
verified
Sandboxing
An agent that can run code can run any code, including code a web page talked it into running.
verified