Skip to main content
affpapa.org
RU EN
skill Claude Code CX Codex Cursor GitHub Copilot official · OpenAI

codex-result-handling — Claude Code skill: handle Codex output

A skill that ensures Codex results are presented to you properly. Preserves findings, inferences, file paths, line numbers, and severity order. Blocks automatic code fixes after review—you must approve changes first.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓ 
npx skills add openai/codex-plugin-cc --skill codex-result-handling -a claude-code
Open on GitHub

What codex-result-handling does

codex-result-handling manages how Codex output reaches you in the IDE. It keeps the structure intact: verdict, summary, findings ordered by severity, inferences, open questions, and affected files. If Codex made edits, the skill lists them; if no issues were found, it says so explicitly.

The skill enforces a hard stop after review. Once Codex reports findings, no file changes happen until you choose which issues to fix. This prevents accidental modifications based on incomplete analysis or inferences marked as uncertain.

If Codex fails or returns an error, the skill shows the most useful stderr output and stops—no guessing, no workarounds. For setup or auth issues, it directs you to /codex:setup.

What it helps with

  • Preserve Codex output structure when sending results to the user
  • Display findings in severity order with exact line numbers
  • Prevent automatic code fixes after review without explicit approval
  • Distinguish facts from inferences and open questions in Codex output
  • Report Codex errors with actionable context instead of attempting workarounds

Example prompts for the agent

  • Run `/codex:review src/auth.ts` and the skill shows findings in the right order with file paths
  • After Codex returns issues, the skill asks which ones you want fixed instead of auto-applying patches
  • If Codex fails to run, the skill prints the error and suggests `/codex:setup`

How to install codex-result-handling

1Skills CLI (Vercel)

npx skills add openai/codex-plugin-cc --skill codex-result-handling -a claude-code

2Manually: put the files in place

~/.claude/skills/codex-result-handling/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/codex-result-handling/

From the same set3

The whole set →
codex-cli-runtime official openai An internal skill that forwards Claude Code tasks to codex-companion through the codex:codex-rescue subagent. Claude Code CX Codex Cursor +1 794installs gpt-5-4-prompting official openai This skill structures Codex and GPT-5.4 prompts for coding, review, diagnosis, and research work. Claude Code CX Codex Cursor +1 726installs

Similar

mentoring-juniors official github This skill mentors junior developers through questions instead of ready-made answers. Use it for learning code, debugging, and getting started with Copilot. GitHub Copilot 6,900installs paseo-committee getpaseo Spawns two agents with contrasting reasoning styles to diagnose a stuck problem, break out of repetitive loops, and synthesize a plan. Useful when you need a fresh perspective and deep root-cause analysis. Claude Code CX Codex Cursor +1 4,700installs transitions-polish Jakubantalik This skill audits existing animations and maps timing, distance, scale, blur, and easing to motion tokens. Use it when motion feels off. Claude Code CX Codex Cursor +1 10,600installs codegen vercel-labs Utilities for generating code from json-render UI trees. Build custom code exporters for Next.js, Remix, and other frameworks, traverse specs, and serialize props. Claude Code CX Codex Cursor +1 8,900installs python-code-style Seth Hobson A skill for Python style rules, linting, formatting, naming, and documentation. Use it when writing code, reviewing changes, or setting project standards. Claude Code CX Codex Cursor 16,456installs firebase-security-rules-auditor Google LLC Audits Firebase rules for data leaks, privilege escalation, and validation gaps. Use it when reviewing Firestore or Cloud Storage access. Claude Code CX Codex Cursor +1 138,200installs improve-codebase-architecture Matt Pocock This skill scans a codebase for architectural friction, then builds a visual HTML report with refactoring candidates. Use it before a deeper architecture review. Claude Code 1,163,035installs tdd Matt Pocock Guides the red-green TDD loop: ensures tests are written first, keep them at public interfaces, and avoid coupling to implementation details. Claude Code 1,141,062installs

Questions about codex-result-handling

How do I install codex-result-handling in Claude Code?

Run npx skills add openai/codex-plugin-cc --skill codex-result-handling -a claude-code, then check ~/.claude/skills/codex-result-handling/SKILL.md.

What does codex-result-handling do?

It ensures Codex results reach you with the right structure: findings ordered by severity, file paths and line numbers preserved, and a clear distinction between facts and uncertain inferences. Most important—it stops automatic fixes and asks for your approval first.

Why doesn't the skill fix issues automatically?

Because Codex output may contain inferences or incomplete analysis. The skill shows you the findings, asks which ones to fix, and only then touches files. You stay in control.

Source text

SKILL.md · Apache-2.0 
---
name: codex-result-handling
description: Internal guidance for presenting Codex helper output back to the user
user-invocable: false
---

# Codex Result Handling

When the helper returns Codex output:
- Preserve the helper's verdict, summary, findings, and next steps structure.
- For review output, present findings first and keep them ordered by severity.
- Use the file paths and line numbers exactly as the helper reports them.
- Preserve evidence boundaries. If Codex marked something as an inference, uncertainty, or follow-up question, keep that distinction.
- Preserve output sections when the prompt asked for them, such as observed facts, inferences, open questions, touched files, or next steps.
- If there are no findings, say that explicitly and keep the residual-risk note brief.
- If Codex made edits, say so explicitly and list the touched files when the helper provides them.
- For `codex:codex-rescue`, do not turn a failed or incomplete Codex run into a Claude-side implementation attempt. Report the failure and stop.
- For `codex:codex-rescue`, if Codex was never successfully invoked, do not generate a substitute answer at all.
- CRITICAL: After presenting review findings, STOP. Do not make any code changes. Do not fix any issues. You MUST explicitly ask the user which issues, if any, they want fixed before touching a single file. Auto-applying fixes from a review is strictly forbidden, even if the fix is obvious.
- If the helper reports malformed output or a failed Codex run, include the most actionable stderr lines and stop there instead of guessing.
- If the helper reports that setup or authentication is required, direct the user to `/codex:setup` and do not improvise alternate auth flows.

© openai · Apache-2.0 · openai/codex-plugin-cc

See also

Project sponsors

Companies that keep this analytics open
Gold sponsor
Nutra.Club

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor