back

Files (1)

• SKILL.md

SKILL.md (first lines)

# back (Codex wrapper)

This is a generated Codex wrapper around the canonical ai-config Claude skill.

Source: [skills/back/SKILL.md](../../skills/back/SKILL.md)

Before acting, read the source skill completely and follow its workflow, adapting it to Codex.

The source lives at `skills/back/SKILL.md` in the same ai-config checkout as this wrapper.
If this wrapper was loaded through `${CODEX_HOME:-$HOME/.codex}/skills/back`, resolve the symlink target for this wrapper directory first, then read `../../skills/back/SKILL.md` relative to that real directory.
Do not resolve that relative path from inside `${CODEX_HOME:-$HOME/.codex}/skills`, because it points back at the wrapper tree.

- Treat `user-invocable` and `allowed-tools` as Claude metadata, not Codex permissions.
- Use the tools available in this Codex session for equivalent operations.
- If the source mentions a Claude-only path such as `~/.claude/skills`, use this repository's `skills/` path while editing.
- Keep procedural changes in the canonical source skill unless the user specifically asks to change this wrapper.

## Tool mappings

The canonical skill names `gh`/`git` commands (and sometimes `mcp__github__*` tools).
Resolve those operations using the full per-model reference at [tool-mappings.md](../../tool-mappings.md), preferring the GitHub MCP tool when this Codex session has it and otherwise using the CLI.

From Morrison-Lab/ai-config, MIT.

More from Morrison-Lab/ai-config