claudekit / updates / claude-code-2-1-202
[ PATCH · ]

Claude Code 2.1.202

Adds a "Dynamic workflow size" setting in `/config` to control how large Claude makes dynamic workflows, plus `workflow.run_id` and `workflow.name` OpenTelemetry attributes so a workflow run can be reconstructed from OTel data. Fixes a batch of Remote Control issues (commands failing with "Unknown command", uncaptioned images and files silently dropped, wrong permission mode), slow session resume in repositories with many git worktrees, and re-invoking an already-loaded skill duplicating its instructions in context. `/review <pr>` is back to a fast single-pass review.

Claude Code
Official announcement →

This article is a summary based on official documentation.

What changed

Claude Code 2.1.202 shipped on July 6, 2026. It adds controls and telemetry for dynamic workflows and clears out regressions across Remote Control (mobile/web), background sessions and agents, and session resume. Network and performance fixes — mTLS certificate rotation, download retries, and faster session resume in repos with many git worktrees — are included too.

New features

  • “Dynamic workflow size” setting in /config

    There was no way to control how large a workflow’s fleet of subagents would get. /config now has a “Dynamic workflow size” setting for controlling how large Claude generally makes dynamic workflows (small/medium/large agent counts). It’s an advisory guideline, not an enforced cap.

  • workflow.run_id and workflow.name OpenTelemetry attributes

    Telemetry emitted by workflow-spawned agents didn’t carry which workflow run it belonged to. workflow.run_id and workflow.name OpenTelemetry attributes are now added to that telemetry, so a workflow run’s activity can be reconstructed from OTel data.

Key improvements

  • /review <pr> back to a fast single-pass review

    /review <pr> had become a slower multi-agent review even when you just wanted a quick pass. It’s now back to a fast single-pass review; use /code-review <level> <pr#> for the multi-agent review at a chosen effort level.

  • /workflows agent list layout

    The agent list was cramped, with narrow titles and a tool-call count on every row. Titles are now wider, there’s a dedicated time column, model names are shorter, and per-row tool-call counts are gone.

  • Clearer MCP error messages

    A server config with url but no type produced a misleading “command: expected string” error. It now shows a clearer error suggesting "type": "http".

Bug fixes

Workflows

  • Workflow scripts with unicode quote escapes in strings being corrupted before parsing — fixed. Workflow parse errors now show the offending line instead of always blaming TypeScript.

Remote Control and mobile/web

  • Commands sent from Remote Control (mobile/web) into an interactive session failing with “Unknown command” — fixed.
  • Images and files sent from the Remote Control mobile or web app without a caption being silently dropped — fixed.
  • /remote-control sessions showing the wrong permission mode in the mobile and web apps — fixed.

Background sessions and agents

  • /rename on background sessions being reverted when the job restarts, which broke addressing the session by its new name — fixed.
  • Opening a chat from claude agents sometimes failing with “currently running as a background agent” followed by a worker crash/respawn loop — fixed.

Session resume and context

  • Resuming a session by name, or opening the resume picker, taking minutes and using a large amount of memory in repositories with many git worktrees — fixed.
  • Re-invoking an already-loaded skill appending a duplicate copy of its instructions to context — fixed.

Network and connectivity

  • Transient mTLS handshake failures when settings were re-applied during an in-place client certificate rotation — fixed.
  • The sign-in URL printed by claude auth login and claude mcp login --no-browser not being reliably clickable when it wraps over SSH — fixed; it is now emitted as a single hyperlink.
  • Installer and updater downloads failing immediately with “aborted” when a proxy or network drops the connection mid-download — fixed; transient connection drops now retry.

Input and UI

  • A crash in the inline Ctrl+R history search when accepting or cancelling while the search was still scanning the history file — fixed.
  • Voice dictation retrying in an unbounded loop when the microphone or audio recorder fails — fixed; repeated capture failures now pause voice input.

Notes

  • Dynamic workflow size is not a hard cap — the small/medium/large setting is an advisory guideline for how large Claude makes workflows, not an enforced limit on agent count.
  • PR review is now split across two commands — use /review <pr> for a fast single-pass review, and /code-review <level> <pr#> when you want a multi-agent review at a chosen effort level.
  • The skill-duplication fix saves context — re-invoking the same skill keeps a single copy of its instructions, cutting wasted context in workflows that reload skills often.