← back NO. #300cea18

NO.: #300cea18
Topic: AGENTS & TOOLS
Source: Claude Code Releases
Published: 2026-05-19 00:48:51
Importance: ★ 8/10 — radar 80
Original: github.com/anthropics/claude-code/releases/tag/v2.1.144

`Claude Code` `v2.1.144` improves background sessions, MCP tools, and terminal stability

Background agents are easier to resume and diagnose. The bigger win is fewer broken long sessions: MCP pagination, terminal corruption, startup hangs, and bad image files now fail less often.

[ KEY POINTS ]

/resume now lists sessions started with claude --bg or agent view, marked as bg; background work is easier to recover after context switching.
Background subagent completion notifications include elapsed time like 3h 2m 5s, useful for spotting expensive or slow automation runs.
/model now changes only the current session; press d in the picker to set defaults, reducing accidental model drift across sessions.
MCP tools/list pagination now returns more than the first page. Tool-heavy setups should stop losing capabilities silently.
Startup hangs when api.anthropic.com is unreachable now time out after 15s, not up to 75s; bad networks hurt less.

Originalgithub.com/anthropics/claude-code/releases/tag/v2.1.144Read original →

// related

#0001
#0001Agents & tools opencode_releases22 hours ago
`opencode` `v1.15.5` adds experimental native OpenAI runtime path
60radar
opencodeOpen-source coding agent — terminal, TUI, and desktop support
The release previews a native OpenAI runtime path and improves resume history with --replay. Worth updating if you use opencode daily; not a major workflow shift yet.
- Experimental native OpenAI runtime path is now behind a flag, making provider integration a little closer to first-class.
- --replay and --replay-limit show recent history when resuming interactive runs, useful for restoring context after interruptions.
- Plugin tools using ask now complete correctly, reducing one of the more annoying failure modes in agent workflows.
- Desktop timelines load faster for large sessions, and stale update downloads are no longer reused.
Source: github.com/anomalyco/opencode/releases/tag/v1.15.5Read original →
FIG-0011:1
60radar
FIG-0011:1
#0002
#0002Agents & tools GitHub Changelogyesterday
GitHub adds REST API auditing for `Copilot` cloud agent repo config
60radar
Repo-level agent settings can now be checked by API instead of manual UI review. Useful for keeping automation permissions visible before cloud agents touch production code.
- New endpoint: Get Copilot cloud agent configuration for a repository, currently in public preview.
- Best fit is policy drift checks across repos: scan whether agent access and configuration match your expected defaults.
- This is governance plumbing, not a coding-speed feature. Worth adopting if Copilot agents run on real repos.
Source: github.blog/changelog/2026-05-18-audit-repository-copiloRead original →
FIG-0021:1
60radar
FIG-0021:1
#0003
#0003Agents & tools GitHub Changelogyesterday
`Copilot Spaces API` is now generally available
70radar
Copilot SpacesGitHub Copilot feature — manages task-specific context spaces
Spaces can now be managed from your own apps via API. Useful for wiring repo context into internal tools or repeatable agent workflows.
- The API supports create, read, update, and delete for Spaces, so context setup no longer has to stay inside GitHub UI.
- Good fit for templates: bootstrap a project space per repo, customer, or feature branch and keep agent context consistent.
- This is more automation surface than end-user feature. Value depends on whether Copilot Spaces is already part of the coding workflow.
Source: github.blog/changelog/2026-05-18-copilot-spaces-api-now-Read original →
FIG-0031:1
70radar
FIG-0031:1

`Claude Code` `v2.1.144` improves background sessions, MCP tools, and terminal stability

// related

`opencode` `v1.15.5` adds experimental native OpenAI runtime path

GitHub adds REST API auditing for `Copilot` cloud agent repo config

`Copilot Spaces API` is now generally available