AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → builders on the Anthropic TypeScript SDK get a beta billing_policy parameter, new billing error codes, and key failure webhooks. Session tool runners also stop executing tools for terminated sessions.
Key takeaways · AI-distilled
Anthropic's TypeScript SDK 1.10.0 adds beta support for a billing_policy request parameter, plus error codes for linked billing refusals, a reached contract spend limit, and unauthorized user-profile billing.
The session tool runner no longer runs tool calls from a terminated session thread or waits for their confirmation. A call it started before reading the termination event still finishes, and is yielded with posted: false if the API refuses the result.
EnvironmentWorker now force-stops work for a terminated session right after retrieving it, before downloading skills or memory stores, and no longer throws SessionMemoryError there. The API reports archived sessions as terminated.
A docs change notes that content_scan.reason on an errored plugin scan explains why there is no verdict and is not a security finding, so code treating any reason as a concern should also check status.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.