What does this skill do, and when should you use it?
This skill gives agents practical test-authoring guidance, with a preference for integration tests when agent logic changes. Those tests belong under core/suite and use test_codex to create a Codex test instance. Unit tests, when necessary, should live in dedicated *_tests.rs files. The guidance also requires identifying major logic changes and user-facing behaviors that need coverage, while encouraging reuse of existing helpers.
It directs the agent to identify major logic changes and user-facing behaviors, prefer integration tests under core/suite using test_codex, place necessary unit tests in dedicated *_tests.rs files, avoid test-only functions in production implementation, and inspect existing helpers to make tests more readable and streamlined.
- A Codex maintainer needs to decide how to test a change to agent logic.
- An engineer has added user-visible behavior and needs corresponding integration coverage.
- A Rust project needs unit tests separated from the main implementation.
- A test suite contains repeated setup and the author needs to look for existing helpers.
How do you install this skill?
- The skill does not specify how to diagnose or report test failures, missing environments, or inability to proceed.
- It omits requirements for sensitive-data handling, isolation, rollback, and user confirmation.
- It lacks versioning, maintenance ownership, changelog, examples, and non-fit scenarios.
The skill is located at .codex/skills/code-review-testing/SKILL.md in the openai/codex repository. The README documents Codex CLI installation, including curl -fsSL https://chatgpt.com/codex/install.sh | sh on macOS/Linux, powershell -ExecutionPolicy ByPass -c "irm https://chatgpt.com/codex/install.ps1 | iex" on Windows, npm install -g @openai/codex, and brew install --cask codex. The README does not document standalone installation of this skill into another client.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/openai/codex.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/.codex/skills/code-review-testing" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Design and implement tests for this agent-logic change, preferring integration tests under core/suite with test_codex; list the major logic changes and user-facing behaviors that need coverage.
Use Codex CLI in the repository and give a request such as: “Design and implement tests for this agent-logic change, preferring integration tests under core/suite with test_codex; list the major logic changes and user-facing behaviors that need coverage.” SKILL.md defines no separate command or trigger syntax.
What are this skill's strengths and limitations?
- Clearly prioritizes integration tests for changes to agent logic.
- Names concrete test locations and conventions, including core/suite, test_codex, and *_tests.rs.
- Discourages test-only production functions and promotes reuse of existing helpers.
- It provides authoring guidance, not test execution or reporting.
- It includes no complete examples, failure-handling workflow, or coverage threshold.
- Its directory and setup conventions are specific to the Codex repository and may need mapping elsewhere.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Code Review and Testing Guidance this page ✓ OpenAI · Official | 43 · Not recommended | ★ 129k | 11d ago | Apache-2.0 |
| RenderCV Testing Context | 61 · Recommended | ★ 18k | 7mo ago | MIT |
| Marmite Development Guide | 58 · Recommended | ★ 886 | 26d ago | AGPL-3.0 |
| Test-Driven Development Workflow | 57 · Use with care | ★ 107k | 12d ago | MIT |
| GitHub Copilot SDK Upgrade Assistant ✓ Microsoft · Official | 28 · Not recommended | ★ 194k | 11d ago | MIT |
How did FollowSkills review this skill?
The skill only gives testing guidance; it executes no commands, requests no privileges, and shows no malware, credential theft, or exfiltration. It does not address sensitive data, external effects, user confirmation, rollback, or test isolation, so points are deducted.
The guidance is internally consistent and identifies integration-test locations, test_codex, unit-test naming, and avoidance of test-only production functions. It lacks abnormal-input handling, dependency-failure guidance, diagnosable errors, and stop conditions; static review also caps the score below 10.
It fits testing decisions for changes to Codex agent logic and provides a reasonably clear scenario. It lacks non-fit boundaries, defined inputs and outputs, edge conditions, semantic trigger rules, and evidence of Chinese or varied-environment support, so points are deducted.
The document is concise and readable, with paths, naming guidance, and a basic progression from integration to unit tests. Repository license and official provenance are available, but the skill has no version, changelog, maintenance owner, installation notes, examples, or troubleshooting, so points are deducted.
It directly guides users toward integration tests or dedicated unit-test files and discourages test-only production helpers. It provides no end-to-end procedure, examples, acceptance output, or comparative-benefit evidence, and static material cannot establish light-review usability, so the score is capped at 7.
The recommendations are auditable against the skill text and repository testing conventions, and the supplied repository contains testing and CI material. There is no skill-specific test suite, third-party execution evidence, or corroboration sufficient for a higher score; nothing was executed.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →