Call Codex
- 380 installs
- 3 repo stars
- Updated January 23, 2026
- dotneet/claude-code-marketplace
call-codex is a Claude Code skill that delegates coding subtasks to the OpenAI Codex CLI via codex exec and merges results back into the active repository session for developers who need a second agent executor.
About
call-codex is a Claude Code marketplace skill from dotneet/claude-code-marketplace that offloads defined coding tasks to OpenAI Codex while Claude Code keeps orchestrating the repository session. The workflow checks whether the codex binary is installed with which codex, then runs codex exec --sandbox danger-full-access with a heredoc task description and recommends maximum Bash timeout for long jobs. Developers reach for call-codex when a subtask benefits from Codex's execution model, when Claude Code should stay in planner-or-reviewer mode, or when parallel agent backends can shorten implementation time. The skill skips subsequent steps cleanly when Codex is not installed.
- Invokes Codex for isolated implementation passes
- Passes repo context and task prompts between agents
- Useful for parallelizing boilerplate or test generation
- Keeps Claude as orchestrator while Codex executes slices
- Marketplace skill with predictable invocation patterns
Call Codex by the numbers
- 380 all-time installs (skills.sh)
- Ranked #2,035 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Jul 29, 2026 (Skillselion catalog sync)
npx skills add https://github.com/dotneet/claude-code-marketplace --skill call-codexAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 380 |
|---|---|
| repo stars | ★ 3 |
| Last updated | January 23, 2026 |
| Repository | dotneet/claude-code-marketplace ↗ |
How do you delegate coding tasks from Claude Code to Codex?
Delegate coding subtasks from Claude Code to OpenAI Codex and merge results back into the active repository session.
Who is it for?
Claude Code users with the Codex CLI installed who want a second agent to execute bounded implementation subtasks.
Skip if: Developers without the codex CLI installed or teams that prohibit danger-full-access sandbox execution on production machines.
When should I use this skill?
A coding subtask should be executed by OpenAI Codex while Claude Code remains the orchestrating session.
What you get
Codex-executed code changes, command output, and merged results in the active git repository session.
- codex task output
- merged repository changes
Files
Call codex to perform a task
Call codex to perform a task.
Checking for the Existence of codex
# Check if codex is installed. Exit code 0 means success, 1 means failure.
# If not installed, skip the subsequent steps of the skill.
which codex
if [ $? -ne 0 ]; then
echo "codex is not installed"
fiRequesting a Task to codex
This task may take a long time. If the timeout value of the Bash tool can be set, please specify the maximum timeout value.
codex exec --sandbox danger-full-access <<EOT
{task_description}
EOTRelated skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10How it compares
Pick call-codex over single-agent Claude Code when a bounded subtask should execute under Codex while Claude retains session orchestration.
FAQ
How does call-codex invoke OpenAI Codex?
call-codex verifies the codex binary with which codex, then runs codex exec --sandbox danger-full-access with the task description passed through a heredoc. Long tasks should use the maximum available Bash timeout.
What happens if Codex is not installed?
call-codex runs which codex first and skips all subsequent delegation steps when the command fails, preventing Claude Code from hanging on a missing OpenAI Codex CLI installation.