Anthropic Claude Code: v2.1.144 Brings /resume for Background Sessions and Fix for 75-Second Hang
Claude Code CLI v2.1.144 introduces /resume support for background sessions showing duration like 'Agent completed · 3h 2m 5s', fixes the 75-second hang on unavailable API, resolves an MCP tools/list pagination bug that silently lost tools, and delivers a range of terminal and MCP fixes.
This article was generated using artificial intelligence from primary sources.
Anthropic released Claude Code CLI v2.1.144 on May 19, 2026, with more than 40 fixes focused on background sessions, reliability under unavailable network conditions, and MCP integrations.
What Is Claude Code and Why Do Background Sessions Matter?
Claude Code is Anthropic’s CLI for running AI agents directly from the terminal. Unlike a web interface, it supports long-running agent tasks — such as refactoring an entire repository or running autonomous tests — that execute in the background while the developer continues other work. Background sessions (launched with claude --bg) were previously not visible in the /resume menu, making it difficult to resume interrupted tasks.
/resume Now Shows Background Sessions — With Duration
The key addition in v2.1.144 is the integration of background sessions into the /resume menu. All sessions launched via claude --bg or the agent view now appear alongside interactive ones, marked with a bg badge. On task completion, the notification includes total elapsed time — e.g. Agent completed · 3h 2m 5s — making it easier to track long automated workflows.
In addition, /bg and left-arrow detach now preserve directories added via /add-dir, and a bug that hid sessions spawned from a background session in the /resume picker has been fixed.
Fix for the 75-Second Hang and MCP Pagination Bug
The previous CLI version could hang for up to 75 seconds at startup if api.anthropic.com was unreachable — which happens behind a VPN, on captive portals, or due to firewall rules. In v2.1.144, secondary API calls now receive a 15-second timeout, effectively eliminating the problem.
The second critical fix concerns MCP servers that return tools across multiple pages (paginated tools/list). Claude Code was silently displaying only the first page — the rest of the tools were lost with no error or warning. Developers using an MCP server with a larger number of tools could unknowingly work with an incomplete set of capabilities.
Other Fixes
The release also brings a range of terminal improvements: self-recovery from corrupted display after window resize without needing Ctrl+L, reduced rendering artifacts in VS Code, and a fix for ghost characters in Windows Terminal with CJK content. MCP images with unsupported MIME types (e.g. SVG) no longer break the conversation but are saved to disk instead.
Frequently Asked Questions
- What is Claude Code?
- Claude Code is Anthropic's official CLI tool that lets developers run AI agents directly from the terminal — supporting interactive and background sessions, MCP server integrations, IDE extensions for VS Code and JetBrains, and agent workflows with long-running tasks.
- What does /resume for background sessions enable?
- The /resume command now displays sessions launched via claude --bg or the agent view alongside interactive ones, labelled with a 'bg' badge. Users can resume any of them, and the duration of completed tasks is shown as a message like 'Agent completed · 3h 2m 5s'.
- What is the MCP tools/list pagination bug and why was it dangerous?
- An MCP (Model Context Protocol) server returning tools across multiple pages (paginated tools/list) was only sending Claude Code the first page — the rest of the tools were silently lost with no error. Developers had no way of knowing tools were missing without manually checking the list — v2.1.144 fixes this.