메뉴
Delegate bounded implementation work to Codex (GPT-5.x) — clear-spec features, migrations, bulk mechanical edits. Use when a task is well-specified enough to hand off whole and worth doing on Codex's near-free usage instead of Claude's, usually in a worktree.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Delegate a Blip file transfer to Codex computer use. Use when the user asks Claude to send, blip, transfer, move, or copy a local file to their phone with Blip; the recipient must be exactly Galaxy S21 Ultra 5G.
Codex (GPT-5.x) computer-use verification — drives the browser or desktop, launches apps, captures screenshots. Use when the user asks to test a flow, verify UI behavior, inspect a running app, capture screenshots, or confirm implemented behavior end-to-end.
Independent second-pass code review from Codex (GPT-5.x). Use when the user wants a second opinion on uncommitted changes, a branch diff, a commit, or a PR, or when a change is broad enough that another agent's perspective is useful before merging.
Drive OpenAI Codex from Claude Code — raw wrapper mechanics for prompts, structured JSON, image generation, and session resume. Use when the user wants to "ask Codex" or delegate a task that codex-review, codex-implementation, and codex-computer-use don't cover; those skills point here for command mechanics.
Record the screen with wrec. Use when the user wants a screen recording, sample video, display/window/app capture, system-audio capture, cursor or Wrec-window visibility control, recording job control, output lookup, or wrec CLI usage.
Blip a local file to Shawn's phone. Use when the user asks to send, blip, transfer, move, or copy an APK or other local file to their phone with Blip; the recipient must be exactly Galaxy S21 Ultra 5G and the transfer must be visually verified.
| name | codex-implementation |
| description | Delegate bounded implementation work to Codex (GPT-5.x) — clear-spec features, migrations, bulk mechanical edits. Use when a task is well-specified enough to hand off whole and worth doing on Codex's near-free usage instead of Claude's, usually in a worktree. |
Codex only gets what you put in the prompt — no conversation context, no memory. Bounded is the contract: a task Codex can complete from the prompt alone, with a done-condition you can check mechanically afterward.
Command mechanics (flags, sandbox, resume) live in the shared wrapper's docs: C:\Users\shawn\.claude\skills\codex\SKILL.md.
Bound the task. Write a self-contained prompt: what to build or change, which files/dirs are in scope, relevant conventions from the codebase, and an explicit done-condition (e.g. "tests in X pass", "every call site of Y migrated"). Done when someone with no context could execute from the prompt alone.
Isolate. If Codex's edits could collide with work in the current tree, create a git worktree and point Codex there. Done when the working root is one Codex can freely mutate.
Run with write access, generous timeout:
node C:\Users\shawn\.claude\skills\codex\scripts\codex-run.mjs --write --cd <worktree-or-repo> "<bounded prompt>"
If it comes back incomplete, resume the same session with a short delta prompt (--resume-last) rather than restarting. Done when Codex reports the done-condition met.
Verify before adopting. Read the full diff, run the tests or checks named in the done-condition yourself, and judge the code against the taste bar — Codex solves problems well but writes TypeScript like a Python dev. Rework or redo with a smarter model anything below the bar; escalating costs less than shipping mediocre work. Done when every changed file has been read and the done-condition independently confirmed.