Sync Codex and Claude Code configs: skills, agents, MCP, permissions

Two situations a lot of us run into right now:

  1. You spent months curating skills, agents, MCP servers, and permissions in ~/.codex (or ~/.claude). You want to try the other one, or move between them. Migrating by hand is slow and easy to get wrong.

  2. You use both side by side. Same skills, same agents, same MCP. Two trees. They drift. Adding anything means two edits.

ai-config-sync-manager treats this as a real sync problem. Diff, plan, apply, backup with rollback. Not a byte copy.

The translation is host-aware:

  • Codex sandbox_mode / web_search / prefix_rule map to Claude tool permissions
  • Codex agent fields (TOML) map to Claude agent frontmatter (YAML), including model alias mapping
  • MCP servers move both ways, including remote MCP with bearer-token env vars
  • Skills sync as folders, not loose files
  • Vocabulary mismatches that can’t auto-translate get paraphrase overrides instead of silent corruption

Zero runtime deps. Single Node ESM CLI. Plugin for both hosts so you can run it from inside either tool.

npx ai-config-sync connect       # register the plugin in both hosts
npx ai-config-sync status        # see what's out of sync
npx ai-config-sync sync          # dry-run plan
npx ai-config-sync sync --apply  # apply with backup

Repo: GitHub - slash9494/ai-config-sync-manager: Bidirectional sync for Claude Code and Codex — instructions, skills, agents, MCP servers, hooks, permissions. · GitHub

Just hit 0.1.0. Curious to hear from anyone running both Codex and Claude Code, especially around vocabulary edge cases the auto-translation should cover.

Welcome to the forum!

For the sake of completeness have to point out this official OpenAI skill. :neutral_face:

Migrate supported instruction files, skills, agents, and MCP config into Codex project and global files.

HolyMoly, I didn’t realise OpenAI offered this feature. I think it would make the repository more robust if we integrated it. Thank you!