[ca] Update CLI versions: Claude Code 2.0.70, Codex 0.73.0

[github-actions]

Update Summary

This update includes 2 CLI version updates published on December 15, 2025.

---

Update Claude Code to 2.0.70

• Previous: 2.0.69 → New: 2.0.70
• Published: December 15, 2025 at 23:49:59 UTC
• Timeline: Single version increment (1 day after 2.0.69)

Changes

New Features:

• Added new delegate permission mode to --permission-mode flag
  • New available modes: acceptEdits, bypassPermissions, default, delegate, dontAsk, plan
  • This enables delegation workflows where Claude Code can delegate decisions to other agents or systems

Bug Fixes: None detected

Breaking Changes: None

Security: None identified

CLI Discovery:

• New command-line option: --permission-mode now accepts delegate as a valid choice
• Help output comparison shows only this single addition

Impact Assessment

• Risk Level: Low
• Affected Features: Permission handling and approval workflows
• Migration: No migration required - backward compatible addition

Package Links

• NPM Package: https://www.npmjs.com/package/`@anthropic-ai/claude-code`
• Specific Release: Version published December 15, 2025

---

Update Codex to 0.73.0

• Previous: 0.72.0 → New: 0.73.0
• Published: December 15, 2025 at 22:39:49 UTC
• Timeline: Single version increment

Release Highlights (from GitHub)

New Features:

• Add ghost snapshot v2 for improved session capture (feat: ghost snapshot v2 openai/codex#8055)
• Support ghost commits in config (feat: config ghost commits openai/codex#7873)
• Reimplement skills loading via SkillsManager and skills/list for consistent discovery (Reimplement skills loading using SkillsManager + skills/list op. openai/codex#7914)
• Add OpenTelemetry tracing for Codex ([codex] add otel tracing openai/codex#7844)

Bug Fixes:

• Prevent panic when session contains a tool call without an output (Do not panic when session contains a tool call without an output openai/codex#8048)
• Avoid triggering keybindings view on input bursts (fix: Don't trigger keybindings view on input burst openai/codex#7980)
• Change default wrap algorithm from OptimalFit to FirstFit (Changed default wrap algorithm from OptimalFit to FirstFit openai/codex#7960)
• Introduce AbsolutePathBuf as part of sandbox config (fix: introduce AbsolutePathBuf as part of sandbox config openai/codex#7856)
• Include Error in log messages (fix: include Error in log message openai/codex#7955)

Merged PRs (from GitHub)

Infrastructure & Dependencies:

• stage new windows sandbox binaries as artifacts openai/codex#8076 - stage new windows sandbox binaries as artifacts
• Fixed formatting issue openai/codex#8069 - Fixed formatting issue
• Update config.md openai/codex#8066 - Update config.md
• chore(deps): bump lru from 0.12.5 to 0.16.2 in /codex-rs openai/codex#8045 - chore(deps): bump lru from 0.12.5 to 0.16.2 in /codex-rs
• chore(deps): bump sentry from 0.34.0 to 0.46.0 in /codex-rs openai/codex#8043 - chore(deps): bump sentry from 0.34.0 to 0.46.0 in /codex-rs
• chore(deps): bump actions/cache from 4 to 5 openai/codex#8039 - chore(deps): bump actions/cache from 4 to 5
• chore(deps): bump actions/download-artifact from 4 to 7 openai/codex#8037 - chore(deps): bump actions/download-artifact from 4 to 7
• chore(deps): bump socket2 from 0.6.0 to 0.6.1 in /codex-rs openai/codex#8046 - chore(deps): bump socket2 from 0.6.0 to 0.6.1 in /codex-rs
• chore(deps): bump actions/upload-artifact from 5 to 6 openai/codex#8038 - chore(deps): bump actions/upload-artifact from 5 to 6

Documentation & Tests:

• chore: fix tooltip typos and align tone openai/codex#8047 - chore: fix tooltip typos and align tone
• docs: document enabling experimental skills openai/codex#8024 - docs: document enabling experimental skills
• docs: update the docs for @openai/codex-shell-tool-mcp openai/codex#7962 - docs: update the docs for @openai/codex-shell-tool-mcp
• docs: remove blanket ban on unsigned integers openai/codex#7957 - docs: remove blanket ban on unsigned integers
• fix: added test helpers for platform-specific paths openai/codex#7954 - fix: added test helpers for platform-specific paths

Code Quality:

• Sync tui2 with tui and keep dual-run glue openai/codex#7965 - Sync tui2 with tui and keep dual-run glue

CLI Discovery

No changes detected in main CLI help output between 0.72.0 and 0.73.0. All changes are internal improvements, bug fixes, and dependency updates.

Impact Assessment

• Risk Level: Low
• Affected Features:
  • Session handling (ghost snapshots, crash prevention)
  • Skills loading system
  • Sandbox configuration
  • Observability (OpenTelemetry tracing)
• Migration: No migration required - all changes are backward compatible improvements

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
• Specific Release: https://github.com/openai/codex/releases/tag/rust-v0.73.0

---

Changes Made

1. Updated pkg/constants/constants.go:

   • DefaultClaudeCodeVersion: 2.0.69 → 2.0.70
   • DefaultCodexVersion: 0.72.0 → 0.73.0

2. Ran make recompile to regenerate all workflow lock files

3. Modified workflow files (39 files updated with new versions)

No Changes Required

The following tools remain at their current versions:

• ✅ Copilot CLI: 0.0.369 (latest: 0.0.369)
• ✅ GitHub MCP Server: v0.24.1 (latest: v0.24.1)
• ✅ Playwright MCP: 0.0.52 (latest: 0.0.52)
• ✅ Playwright Browser: v1.57.0 (latest: v1.57.0)

---

Overall Assessment

Both updates are low-risk, backward-compatible improvements:

• Claude Code 2.0.70: Adds optional delegate permission mode for enhanced workflow flexibility
• Codex 0.73.0: Internal improvements including better session handling, skills loading, crash prevention, and observability

Recommendation: Safe to deploy immediately. No breaking changes or migration steps required.

AI generated by CLI Version Checker