The runtime associates the result with the original call and adds it to the next model context. Because tool output may be incomplete, malicious, or stale, the agent should treat it as evidence rather than trusted instruction.
A tool result is the structured or textual output returned to an agent after an external operation runs.
The runtime associates the result with the original call and adds it to the next model context. Because tool output may be incomplete, malicious, or stale, the agent should treat it as evidence rather than trusted instruction.