[ca] CLI Version Updates Available: Copilot CLI 0.0.409 and Codex 0.101.0

[github-actions]

Summary

Two CLI tools have new versions available with bug fixes, new features, and improvements:

• Copilot CLI: 0.0.407 → 0.0.409 (2 versions)
• Codex: 0.99.0 → 0.101.0 (2 versions)

Status: ✅ constants.go has been updated with new versions
Next Step: Run integration tests to verify compatibility before merging

---

GitHub Copilot CLI

Version Timeline

• Previous: 0.0.407
• Intermediate: 0.0.408 (February 12, 2026)
• Latest: 0.0.409 (February 12, 2026)
• Release Frequency: Both versions released same day

Breaking Changes

None identified

Key Features (0.0.409)

• VS Code Integration: New /ide command enables VS Code integration
• Quick Help Overlay: Press ? key for grouped keyboard shortcuts
• Scrollable Permission Prompts: Permission prompts with extensive diffs now scrollable in alt-screen mode
• Plugin Discovery: Default plugin marketplaces (copilot-plugins, awesome-copilot) added
• GitHub MCP Enhancement: Added list_copilot_spaces tool to default GitHub MCP config

Key Features (0.0.408)

• Streamer Mode: New /streamer-mode command hides preview model names and quota details during streaming
• Mouse Support: Text selection with mouse in --alt-screen mode
• Slash Command Improvements: Substring matching in slash command autocomplete
• Keyboard Shortcut Change: Run command shortcut changed from ctrl+p to ctrl+s

View Full Changelog

Version 0.0.409 (February 12, 2026)

Features

• /diff uses full screen in alt-screen mode
• Quick help overlay accessible via ? key with grouped shortcuts
• Theme preview positioning improved for screen reader accessibility
• Added list_copilot_spaces tool to default GitHub MCP config
• Subagents now deliver comprehensive responses
• VS Code integration enabled via /ide command
• Permission prompts with extensive diffs became scrollable in alt-screen
• Default plugin marketplaces (copilot-plugins, awesome-copilot) added

Impact on gh-aw

• New GitHub MCP tool list_copilot_spaces may be useful for workspace management features
• VS Code integration provides new interaction model for developer workflows
• Enhanced plugin marketplace improves extensibility

Version 0.0.408 (February 12, 2026)

Features

• /streamer-mode to hide preview model names and quota details for streaming
• Enhanced shellId flexibility to prevent errors when numeric values passed
• Background task hints refresh when detached shells are terminated
• Mouse text selection functionality added to --alt-screen mode
• Substring matching integrated into slash command autocomplete

Bug Fixes

• Commands with extensive output no longer cause crashes
• Fixed duplicate/ghost lines appearing when resizing terminal in alt-screen
• MCP servers now respect the cwd working directory property

Keyboard Shortcut Changes

• Run command shortcut changed from ctrl+p to ctrl+s

Impact on gh-aw

• MCP cwd property fix may improve file path resolution in workflows
• Crash fixes for extensive output improve stability for long-running agent tasks
• Terminal rendering improvements benefit alt-screen usage patterns

View Migration Guide

Migration Steps

1. Update constants.go (✅ Completed)

   • Updated DefaultCopilotVersion from 0.0.407 to 0.0.409

2. Run Integration Tests (⚠️ Required)

   • Execute full integration test suite per the warning in constants.go
   • Verify agent execution works with new version
   • Test MCP server interactions with cwd property fix
   • Validate no regressions in workflow execution

3. Test New Features (Optional)

   • Test /ide command for VS Code integration compatibility
   • Verify list_copilot_spaces tool works if GitHub MCP config is affected
   • Check if alt-screen mode improvements affect terminal-based workflows

4. Recompile Workflows (Required)

   • Run make recompile to update workflow lock files
   • Commit updated pkg/constants/constants.go and generated workflow files

Breaking Changes

None expected. All changes are additive or fixes.

Rollback Plan

If issues arise, revert DefaultCopilotVersion to 0.0.407 and run make recompile.

Package Links

• NPM Package: https://www.npmjs.com/package/`@github/copilot`
• Repository: https://github.com/github/copilot-cli (may be private)

---

OpenAI Codex CLI

Version Timeline

• Previous: 0.99.0
• Intermediate: 0.100.0 (February 12, 2026)
• Latest: 0.101.0 (February 12, 2026)
• Release Frequency: Both versions released same day

Breaking Changes

None identified

Key Features (0.101.0)

• Model Stability: Model resolution now preserves requested model slug
• Memory Management: Developer messages excluded from phase-1 memory input to reduce noise
• Performance: Memory phase processing concurrency reduced for better stability under load

Key Features (0.100.0)

• JavaScript REPL Runtime: Experimental feature-gated js_repl runtime with persistent state across tool calls
• Multiple Rate Limits: Support for multiple simultaneous rate limits across protocol, backend, and TUI
• Websocket Transport: Reintroduced app-server websocket transport with split inbound/outbound architecture
• Memory Management: New TUI slash commands /m_update and /m_drop for memory management
• Sandbox Capabilities: Promoted sandbox capabilities on Linux and Windows; new ReadOnlyAccess policy shape
• Apps SDK Support: ChatGPT connector now handles Apps SDK apps

View Full Changelog

Version 0.101.0 (February 12, 2026)

Bug Fixes

• Model resolution now preserves the requested model slug when selecting by prefix (fix(core) model_info preserves slug openai/codex#11602)
• Developer messages are now excluded from phase-1 memory input, reducing noise (exclude developer messages from phase-1 memory input openai/codex#11608)
• Memory phase processing concurrency reduced for stability under load (chore: reduce concurrency of memories openai/codex#11614)

Chores

• Cleaned and simplified phase-1 memory pipeline code paths (chore: drop and clean from phase 1 openai/codex#11605)
• Repository maintenance: formatting and test-suite hygiene updates (fix: fmt openai/codex#11619)

Merged PRs

• chore: drop and clean from phase 1 openai/codex#11605: chore: drop and clean from phase 1
• fix(core) model_info preserves slug openai/codex#11602: fix(core) model_info preserves slug
• exclude developer messages from phase-1 memory input openai/codex#11608: exclude developer messages from phase-1 memory input
• Add cwd to memory files openai/codex#11591: Add cwd to memory files
• chore: reduce concurrency of memories openai/codex#11614: chore: reduce concurrency of memories
• fix: fmt openai/codex#11619: fix: fmt

Full Changelog: openai/codex@rust-v0.100.0...rust-v0.101.0

Version 0.100.0 (February 12, 2026)

New Features

• Experimental JavaScript REPL runtime (js_repl) with persistent state across tool calls (Add feature-gated freeform js_repl core runtime openai/codex#10674)
• Support for multiple simultaneous rate limits across protocol, backend client, and TUI (feat: support multiple rate limits openai/codex#11260)
• Reintroduced app-server websocket transport with split inbound/outbound architecture (Reapply "Add app-server transport layer with websocket support" openai/codex#11370, app-server: thread resume subscriptions openai/codex#11474)
• Memory management slash commands /m_update, /m_drop with expanded memory-read/metrics (feat: mem slash commands openai/codex#11569, feat: memory read path openai/codex#11459, feat: metrics to memories openai/codex#11593)
• Apps SDK apps support in ChatGPT connector ([apps] Allow Apps SDK apps. openai/codex#11486)
• Promoted Linux bubblewrap and Windows sandbox capabilities; new ReadOnlyAccess policy shape (feat(core): promote Linux bubblewrap sandbox to Experimental openai/codex#11381, Promote Windows Sandbox openai/codex#11341, feat: make sandbox read access configurable with ReadOnlyAccess openai/codex#11387)

Bug Fixes

• Fixed websocket incremental output duplication and completion handling (Do not resend output items in incremental websockets connections openai/codex#11383, Do not attempt to append after response.completed openai/codex#11402, Handle response.incomplete openai/codex#11558)
• Improved websocket session stability during reconnects (Pump pings openai/codex#11413, Hide the first websocket retry openai/codex#11548)
• Fixed stale thread entries by dropping missing rollout files (Ensure list_threads drops stale rollout files openai/codex#11572)
• Fixed Windows multi-line paste reliability in terminals, especially VS Code (fix(tui): increase paste burst char interval on Windows to 30ms openai/codex#9348)
• Fixed rate-limit limit_name inheritance issue (fix: stop inheriting rate-limit limit_name openai/codex#11557)
• Reduced repeated skill parse-error spam by increasing file-watcher debounce from 1s to 10s (Increased file watcher debounce duration from 1s to 10s openai/codex#11494)

Documentation

• Added JS REPL documentation and config/schema guidance (Add feature-gated freeform js_repl core runtime openai/codex#10674)
• Updated app-server websocket transport documentation (Reapply "Add app-server transport layer with websocket support" openai/codex#11370)

Chores

• Split codex-common into focused codex-utils-* crates (feat: split codex-common into smaller utils crates openai/codex#11422)
• Improved Rust release pipeline for Windows and musl targets (feat: use more powerful machines for building Windows releases openai/codex#11488, feat: build windows support binaries in parallel openai/codex#11500, Fix rust-release failures in musl linking and release asset upload openai/codex#11556)
• Prevented GitHub release asset upload collisions (rust-release: exclude cargo-timing.html from release assets openai/codex#11564)

Full Changelog: openai/codex@rust-v0.99.0...rust-v0.100.0

View Migration Guide

Migration Steps

1. Update constants.go (✅ Completed)

   • Updated DefaultCodexVersion from 0.99.0 to 0.101.0

2. Test Core Functionality (Recommended)

   • Verify agent execution with Codex engine
   • Test model resolution and selection
   • Validate memory management doesn't introduce issues
   • Check rate-limit handling if workflows use rate-limited operations

3. Evaluate New Features (Optional)

   • Consider JavaScript REPL runtime (js_repl) for JS/TS workflow tasks
   • Review new sandbox ReadOnlyAccess policy for security configurations
   • Explore memory management commands /m_update, /m_drop for debugging

4. Recompile Workflows (Required)

   • Run make recompile to update workflow lock files
   • Commit updated pkg/constants/constants.go and generated workflow files

Breaking Changes

None expected. All major changes are feature-gated or additive.

Rollback Plan

If issues arise, revert DefaultCodexVersion to 0.99.0 and run make recompile.

Package Links

• NPM Package: https://www.npmjs.com/package/`@openai/codex`
• Repository: https://github.com/openai/codex
• Release Notes: https://github.com/openai/codex/releases
• 0.101.0 Release: https://github.com/openai/codex/releases/tag/rust-v0.101.0
• 0.100.0 Release: https://github.com/openai/codex/releases/tag/rust-v0.100.0

---

Impact Assessment

Copilot CLI (0.0.407 → 0.0.409)

• Risk: Low
• Affects: Agent execution, MCP server interactions, terminal UI
• Benefits:
  • MCP cwd property fix improves file path handling
  • Crash fixes for extensive output improve long-running task stability
  • New GitHub MCP tool provides workspace management capabilities

Codex (0.99.0 → 0.101.0)

• Risk: Low
• Affects: Codex engine execution, model resolution, memory management
• Benefits:
  • Model slug stability prevents unexpected model switching
  • Memory management improvements reduce noise and improve performance
  • Websocket stability fixes benefit real-time agent interactions
  • Sandbox enhancements provide better security controls

Recommendations

1. Priority: Medium - Update at next convenient maintenance window
2. Testing Strategy:
   • Run full integration test suite for Copilot CLI (required per warning)
   • Test Codex engine with various workflows to verify model resolution
   • Validate MCP server cwd property works correctly
3. Rollout Plan:
   • Update in development environment first
   • Monitor for any regressions over 24-48 hours
   • Deploy to production if no issues detected
4. Notes:
   • Both tools released same day with coordinated improvements
   • No breaking changes expected, all updates are additive or fixes
   • Memory management and stability improvements align well with gh-aw use cases

---

References:

• Workflow Run: §21971666865

AI generated by CLI Version Checker

• expires on Feb 15, 2026, 2:01 AM UTC