30+ fixes across multi-agent reliability—stacked skills, SSL clarity, streaming durability, error propagation all tightened
Claude Code v2.1.199 — Multi-skill invocations and comprehensive bug fixes
Original: v2.1.199
Importance: Claude Code 運用ユーザーに直結する安定性向上とマルチエージェント周りの重大バグ修正が30項目超
Summary
Anthropic released Claude Code v2.1.199, improving multi-slash-skill invocation to load up to 5 leading skills simultaneously rather than just the first. The release fixes 30+ bugs spanning SSL/TLS certificate handling, streaming response loss after partial output, subagent error propagation, Linux daemon crashes after unclean shutdown, SSH+macOS cold-start failures, rate-limit retry behavior, and session transcript bloat. Background agent stability and error reporting throughout the multi-agent architecture have been substantially enhanced.
Key Points
- Multi-skill invocation improved; up to 5 leading skills now load (was 1)
- SSL/TLS errors fail immediately with actionable guidance; no wasted retries
- Partial streaming output preserved when mid-stream API errors occur
- Subagent errors now correctly propagated to parent agent
- Linux daemon crash loop fixed; macOS SSH cold-start regression (v2.1.196) resolved
Source: https://github.com/anthropics/claude-code/releases/tag/v2.1.199
Outlet: Claude Code Releases
This article is an AI-generated summary (OpenAI GPT-4o-mini) of publicly available information from Anthropic, OpenAI, Google, Meta, Mistral, DeepSeek, Sakana, and other vendors. The original source URL is always provided in accordance with fair-use citation requirements. Summaries are AI-generated and may contain mistranslations or misinterpretations. Always verify details with the original source.