SNACK three-line summary
- Claude Code v2.1.139 added agent view for seeing multiple work sessions and /goal for setting completion conditions.
- Claude Code v2.1.140 followed with fixes for /goal, background service behavior, settings hot-reload, /loop wakeups, and Windows event-loop stalls.
- The update shows AI coding tools moving from simple conversational helpers toward managed agent workflows.
Screenshots and video links
The translated article uses the same screenshots, embeds, and attached video links as the Korean original.
Snackgirls editor note
- Nea — The feature sounds convenient, but it also makes permissions, cost visibility, and human approval more important.
- Red — The point is not one more line of code. It is that coding AI is starting to look like a workroom with status boards.
- Kirari🌟 — Seeing running, waiting, and completed sessions in one place makes AI work feel less invisible and easier to follow.
Release timing
According to official GitHub releases, Claude Code v2.1.139 was released on May 11, 2026, and v2.1.140 followed on May 12, 2026. The article treats the update as fresh AI agent news checked on May 13, 2026.
What Claude Code is
Claude Code is Anthropic’s agentic coding system. Rather than only suggesting code, it can read a codebase, edit multiple files, run commands, assist with git workflows, and help carry out development tasks inside a project folder.
Agent view
The notable v2.1.139 addition is agent view. The release describes a list of Claude Code sessions, including sessions that are running, waiting for user input, or completed. Users can start it with the claude agents command. It is presented as a Research Preview, but the direction is clear: multiple AI work sessions need visibility.
/goal
The /goal command lets users set completion conditions so Claude continues working toward that target. The release says it works in interactive mode, -p print mode, and Remote Control, with an overlay for elapsed time, turns, and token count. The value is keeping a long coding job oriented around the intended finish line.
Why v2.1.140 matters
The next release focused on stability: fixing a case where /goal could stop silently with certain hook settings, improving claude –bg background service behavior, addressing a settings hot-reload regression, fixing unnecessary /loop wakeups, and reducing repeated Windows event-loop stalls. For agents, reliability is part of usefulness.
How this compares with Hermes Agent
The article compares the broader direction with Hermes Agent v0.13.0, which also emphasized Kanban, /goal, gateway recovery, and security. The two tools are different, but both show the same trend: agent tools need goals, visible status, recovery, permissions, and human oversight.
Game Sunakku reading
This does not mean AI replaces developers. It means the operator role changes: define the goal, limit authority, watch cost and token use, verify results, and step in when a session waits or stalls. Agent view and /goal make that structure easier to manage.
Feature table
| Feature or fix | Version | Practical meaning |
|---|---|---|
| agent view | v2.1.139 | Shows running, waiting, and completed Claude Code sessions |
| /goal | v2.1.139 | Sets a completion condition for longer agent work |
| Background service fixes | v2.1.140 | Helps longer-running operation behave more reliably |
| Settings hot-reload and /loop fixes | v2.1.140 | Reduces operational surprises during active use |
| Windows event-loop stall fixes | v2.1.140 | Improves reliability in Windows environments |
Sources and check date · Based on the original Game Sunakku article. Checked: June 6, 2026
Leave a comment