[ca] CLI Version Updates: Claude Code 2.1.105, Copilot 1.0.25, Codex 0.120.0, Gemini 0.37.2

[github-actions]

Summary

4 CLI tools have been updated in pkg/constants/version_constants.go and workflows recompiled.

Tool	Previous	New	Published
Claude Code	2.1.98	2.1.105	2026-04-13
Copilot CLI	1.0.21	1.0.25	2026-04-13
Codex	0.118.0	0.120.0	2026-04-11
Gemini CLI	0.37.1	0.37.2	2026-04-13
GitHub MCP Server	v0.32.0	v0.32.0	— no change —
Playwright MCP	0.0.70	0.0.70	— no change —
Playwright Browser	v1.59.1	v1.59.1	— no change —
MCP Gateway	v0.2.19	v0.2.19	— no change —

---

Update: Claude Code (2.1.98 → 2.1.105)

• Intermediate versions: 2.1.100, 2.1.101, 2.1.104
• Release timeline: 2.1.98 → 2.1.100 → 2.1.101 → 2.1.104 → 2.1.105 (all released ~Apr 7–13)
• Breaking changes: None detected
• CLI changes: No visible changes in --help output vs prior version

View CLI Discovery

Compared help output of 2.1.101 (last cached) vs 2.1.105. No additions or removals detected in the command/flag set.

Key flags available: --bare, --permission-mode, --json-schema, --worktree, --tmux, --plugin-dir, --agents, --fork-session, --from-pr, --mcp-config, --strict-mcp-config, and more.

Impact Assessment

• Risk: Low
• Affects: All Claude-engine agentic workflows
• NPM Package: https://www.npmjs.com/package/`@anthropic-ai/claude-code`

---

Update: GitHub Copilot CLI (1.0.21 → 1.0.25)

• Intermediate versions: 1.0.22, 1.0.23, 1.0.24
• Breaking changes: None detected
• Repository: Private — release notes not accessible

View CLI Changes (1.0.25 vs 1.0.24)

New flags added in 1.0.25:

• --remote — Export session events to Mission Control and enable remote steering
• --no-remote — Disable remote session export and steering

These flags enable integration with a remote Mission Control dashboard for monitoring and steering agentic sessions.

Subcommand Help Changes

• config help topic: Structure unchanged
• environment help topic: No changes

View Config Settings (copilot help config)

allowed_urls, autoUpdate, banner, bashEnv, beep, powershellFlags, compactPaste, copyOnSelect, custom_agents.default_local_only, denied_urls, experimental, includeCoAuthoredBy, companyAnnouncements

View Environment Variables (copilot help environment)

COPILOT_ALLOW_ALL, COPILOT_AUTO_UPDATE, COPILOT_CUSTOM_INSTRUCTIONS_DIRS, COPILOT_EDITOR, COPILOT_GITHUB_TOKEN, COPILOT_HOME, COPILOT_MODEL, COPILOT_OFFLINE, COPILOT_PROVIDER_BASE_URL/TYPE/API_KEY/BEARER_TOKEN/WIRE_API, GH_HOST, HTTP_PROXY, NO_COLOR, NO_PROXY, PLAIN_DIFF

Impact Assessment

• Risk: Low-Medium (verify --no-ask-user still functional, MCPs load correctly)
• Affects: All Copilot-engine agentic workflows
• NPM Package: https://www.npmjs.com/package/`@github/copilot`

---

Update: OpenAI Codex (0.118.0 → 0.120.0)

• Intermediate versions: 0.119.0 (2026-04-10), 0.120.0 (2026-04-11)

View 0.120.0 Highlights (from GitHub release)

New Features

• Realtime V2 can now stream background agent progress while work is still running and queue follow-up responses until the active response completes (Stream Realtime V2 background agent progress openai/codex#17264, Queue Realtime V2 response.create while active openai/codex#17306)
• Hook activity in the TUI is easier to scan, with live running hooks shown separately and completed hook output kept only when useful ([codex] Improve hook status rendering openai/codex#17266)
• Custom TUI status lines can include the renamed thread title (Add thread title to configurable TUI status line openai/codex#17187)
• Code-mode tool declarations now include MCP outputSchema details so structured tool results are typed more precisely (Add output_schema to code mode render openai/codex#17210)
• SessionStart hooks can distinguish sessions created by /clear from fresh startup or resume sessions (Support clear SessionStart source openai/codex#17073)

Bug Fixes

• Fixed Windows elevated sandbox handling for split filesystem policies (fix: support split carveouts in windows elevated sandbox openai/codex#14568)
• Fixed sandbox permission handling for symlinked writable roots and carveouts (fix(permissions): fix symlinked writable roots in sandbox permissions openai/codex#15981)
• Fixed codex --remote wss://... panics by installing Rustls crypto provider before TLS websocket connections (Install rustls provider for remote websocket client openai/codex#17288)
• Preserved tool search result ordering instead of alphabetically reordering results (preserve search results order in tool_search_output openai/codex#17263)
• Fixed live Stop-hook prompts so they appear immediately (Emit live hook prompts before raw-event filtering openai/codex#17189)

Release Notes: https://github.com/openai/codex/releases/tag/rust-v0.120.0

View 0.119.0 Highlights (from GitHub release)

New Features

• Realtime voice sessions now default to the v2 WebRTC path with configurable transport and native TUI media support
• MCP Apps and custom MCP servers gained richer support including resource reads, tool-call metadata, and server-driven elicitations
• Remote/app-server workflows now support egress websocket transport, remote --cd forwarding, and an experimental codex exec-server subcommand
• The TUI can copy the latest agent response with Ctrl+O
• /resume can now jump directly to a session by ID or name from the TUI

Bug Fixes

• TUI starts faster by fetching rate limits asynchronously
• Resume flows are more stable with improved picker behavior
• Composer and chat behavior improvements including fixed paste teardown and CJK word navigation

Release Notes: https://github.com/openai/codex/releases/tag/rust-v0.119.0

Impact Assessment

• Risk: Low
• Affects: All Codex-engine agentic workflows
• NPM Package: https://www.npmjs.com/package/`@openai/codex`

---

Update: Google Gemini CLI (0.37.1 → 0.37.2)

• Breaking changes: None detected
• Repository: No GitHub release notes available
• Published: 2026-04-13T20:13:51Z

Impact Assessment

• Risk: Low
• Affects: All Gemini-engine agentic workflows
• NPM Package: https://www.npmjs.com/package/`@google/gemini-cli`

---

Recommendations

1. Validate Copilot 1.0.25: The new --remote/--no-remote flags should not affect non-interactive agentic runs. Verify that --no-ask-user flag (added in 1.0.19) remains functional.
2. Codex improvements: SessionStart hook enhancement (Support clear SessionStart source openai/codex#17073) enables better session lifecycle tracking — note the new /clear source type if using session-start hooks.
3. All tools: Run standard agentic workflow smoke tests to confirm tool loading and model routing work end-to-end.

References:

• §24380814790

Generated by CLI Version Checker · ● 395.5K · ◷

• expires on Apr 16, 2026, 4:31 AM UTC

[pelikhan]

@copilot do not modify copilot version