← All Claude Code issues

This week in Claude Code · May 10, 2026

Claude Agent SDK Python: Seven Releases Bring Hook Streaming, API Error Codes, and CLI Bumps

Get the next issue in your inbox. Weekly · Free

unsubscribe anytime.

TLDR;

The Claude Agent SDK for Python has shipped seven releases in quick succession (v0.1.74 through v0.1.80). The headline additions are hook event streaming via the new includeHookEvents option, deferred tool-use round trips, and an api_error_status field on ResultMessage that surfaces HTTP error codes like 429 and 529 without leaking sensitive data. Bug fixes include actionable error messages replacing the generic exit-code exception and a PermissionUpdate deserialization fix. Versions 0.1.75 through 0.1.80 also carry bundled CLI updates from 2.1.131 to 2.1.138. Upgrade to v0.1.80 now to stay current with all CLI and SDK improvements.

Agent SDK & Managed Agents

v0.1.74: Hook Event Streaming and Deferred Tool-Use Support Added

  • The new include_hook_events option on ClaudeAgentOptions causes PreToolUse, PostToolUse, Stop, and other hook events to be yielded from the message stream as HookEventMessage objects, matching TypeScript SDK behavior.
  • The same release adds support for the 'defer' hook decision and a new DeferredToolUse dataclass on ResultMessage, completing deferred tool-use round-trip parity with the TypeScript SDK.

v0.1.76: api_error_status Field and Permission Deserialization Fix

  • ResultMessage now exposes api_error_status (int or None) so callers can safely classify API failures such as rate-limit (429) or overload (529) errors without inspecting raw error text.
  • A separate bug fix ensures ToolPermissionContext.suggestions returns properly typed PermissionUpdate instances rather than raw dicts.

v0.1.77: Actionable Error Messages and Skill Option Deprecation

  • The generic 'Command failed with exit code 1' exception after an error result is replaced with one carrying the result's actual error text (for example, 'Reached maximum number of turns'), aligning with TypeScript SDK behavior.
  • The 'Skill' string in allowed_tools is now deprecated in favor of the skills option on ClaudeAgentOptions for more granular skill control.

Action Items

Immediate

  • Upgrade to claude-agent-sdk==0.1.80 to pick up CLI 2.1.138 and all bug fixes from the v0.1.74 to v0.1.80 wave.
  • Migrate any allowed_tools 'Skill' strings to the skills option on ClaudeAgentOptions before the deprecated path is removed.
  • Add api_error_status inspection to your error-handling logic to distinguish rate-limit (429) from overload (529) failures without parsing error text.
  • Audit any pre-tool-use permission hooks that read ToolPermissionContext.suggestions to ensure they handle PermissionUpdate instances rather than raw dicts after the v0.1.76 deserialization fix.
  • Enable include_hook_events on ClaudeAgentOptions if your agent needs to observe or log PreToolUse, PostToolUse, or Stop lifecycle events from the stream.

All Resources

Every Monday morning

This week in Claude Code

Five minutes, every Monday — the tools, releases and tactics for shipping solo.

unsubscribe anytime.

This week for builders

Five minutes, every Monday — the tools, releases and tactics for shipping solo.

unsubscribe anytime.