claude-agent-sdk-typescript v0.3.318
- Source
- github.com
- Author
- ashwin-antTop Viber
- Date

Context
Version 0.3.318 of Anthropic's TypeScript SDK mostly gives apps built on it better signals about why a session might stop working (release notes). Result messages gain an optional `credential_expires_at` field giving the date the stored claude.ai sign-in ends, and startup can now fail with a `config_file_invalid` reason when a configuration file read at startup is not valid JSON. Both mirror Claude Code 2.1.318, which the SDK now matches, where the same release made sign-in expiry visible and stopped non-interactive launches from overwriting a broken config file with defaults.
The smaller changes: interrupting a session now also lists, cancels and stops a message that was queued behind a model change; progress summaries for a waiting on a likely-cold prompt cache can wait up to five minutes and are skipped if the subagent finishes first; and `status` in `PushNotificationInput` is now optional. Code that handles specific startup failure reasons may want to account for the new one, and TypeScript code that reads `status` from `PushNotificationInput` should allow for it to be missing.
- 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.
- multi-agent — Using several AI agents on one problem — splitting work in parallel, checking each other, or filling different roles like planner and reviewer.
Checking sign-in…
Loading comments…





