Scoutari

AI builder 的決策情報台。

瀏覽

  • 今日情報
  • Skill 榜單
  • AI 週報
  • 關於與編輯方針
  • RSS

主題

  • 模型發布
  • API 與價格
  • MCP 與 Skills
  • AI Coding
  • 中國模型
  • 開源模型
  • Agent
  • Infra / 成本
  • 值得試用
© 2026 Scoutari · scoutari.com摘要由 AI 輔助生成,並始終附上原文連結。
情報模型收藏週報設定
Tuesday, July 7, 2026Scoutari
Log inSign up
Sign up
ModelsSkill RadarSavedPersonalizeSubmit sourceLog in
IntelModelsSkill RadarWeeklySavedPersonalizeSubmit source
Back to timeline

Tue, July 706:51ToolsModel releasesMCP & SkillsAI coding

Claude Code v2.1.202: Dynamic Workflow Size & OpenTelemetry Attributes

View originalOpen Claude Code

Decision Brief

What changedClaude Code v2.1.202 adds dynamic workflow size settings and OpenTelemetry attributes for workflow agents, alongside multiple bug fixes.
Why it mattersDevelopers automating long workflows can now control agent count via dynamic workflow size (small/medium/large) and reconstruct workflow activity from OTel data, improving observability and flexibility.
Who should careAI coding tool users, Agent builders
Affected stackClaude CodeMCP
Builder actionUpgrade Claude Code to v2.1.202
Source confidenceHigh · Official release / blog / repo

v2.1.202 introduces a 'Dynamic workflow size' setting in /config, letting users choose small/medium/large as a suggestion for agent count in dynamic workflows (not a hard limit). It also adds workflow.run_id and workflow.name OpenTelemetry attributes to workflow-generated agents, enabling workflow activity reconstruction from OTel data. Bug fixes include: crash on accept/cancel during Ctrl+R history search, /rename in background sessions reverting after task restart, brief handshake failures during mTLS certificate rotation, ignored Remote Control commands and untitled images/files, non-clickable login URLs in SSH for Claude auth/mcp login, chat opening from agent failing due to background agent conflicts, unicode quote escaping errors in workflow scripts, infinite microphone retry on failure, wrong permission mode in remote control, slow session restore in repos with many git worktrees, installer abort on network interruption, and duplicate commands from reloading already-loaded skills. /workflows agent list UI improved: wider titles, dedicated time column, shorter model names, removed per-line tool call count. MCP error messages improved: missing type field now suggests 'type: http' instead of misleading 'command: expected string'. /review reverted to single quick review; use /code-review for multi-agent reviews. Developers should note these changes to optimize workflow configuration and debugging.

Summary basis: official / RSS sourceCompiled from the source scope noted above; the original remains authoritative.

Sources

  • Claude Code(GitHub Releases)

    Claude Code release notes: new features, hooks, skills, and agent capabilities.

  • Claude Code(GitHub Releases)

Related intel

  • Headroom: Open-Source Tool to Compress Tool Outputs, Logs, Files, and RAG Chunks by 60-95% Tokens Without Quality Loss

    Headroom is an open-source project that compresses tool outputs, logs, files, and RAG chunks before sending them to LLMs, reducing token usage by 60-95% while maintaining answer quality.

  • Anthropic Launches Official Claude Code Plugin Directory

    Anthropics/claude-plugins-official is trending on GitHub, providing an official directory for Claude Code plugins.

  • Serena: MCP Toolkit for Semantic Retrieval and Editing of Coding Agents

    Serena is an open-source MCP toolkit designed for coding tasks, providing semantic search and editing capabilities.

  • DeusData/codebase-memory-mcp: High-Performance Code Intelligence MCP Server Indexes Codebase as Persistent Knowledge Graph

    High-performance MCP server indexing codebases into persistent knowledge graphs with millisecond processing for average repos.