← 返回 Skills 市场

Exunit Code Review

Name: Exunit Code Review
Author: anderskev

作者 Kevin Anderson · GitHub ↗ · v1.2.1 · MIT-0

cross-platform ✓ 安全检测通过

168

总下载

当前安装

版本数

在 OpenClaw 中安装

/install exunit-code-review

功能描述

Reviews ExUnit test code for proper patterns, boundary mocking with Mox, and test adapter usage. Use when reviewing _test.exs files or test helper configurat...

使用说明 (SKILL.md)

ExUnit Code Review

Quick Reference

Issue Type	Reference
Async tests, setup, describe, tags	references/exunit-patterns.md
Behavior-based mocking, expectations	references/mox-boundaries.md
Bypass, Swoosh, Oban testing	references/test-adapters.md
What to mock vs real, Ecto sandbox	references/integration-tests.md

Mock Boundary Philosophy

Mock at external boundaries:

HTTP clients, external APIs, third-party services
Slow resources: file system, email, job queues
Non-deterministic: DateTime.utc_now(), :rand

DO NOT mock internal code:

Contexts, schemas, GenServers
Internal modules, PubSub
Anything you wrote

Review Checklist

Test Structure

Tests are async: true unless sharing database state
Describe-blocks group related tests
Setup extracts common test data
Tests have clear arrange/act/assert structure

Mocking

Mox used for external boundaries (HTTP, APIs)
Behaviors defined for mockable interfaces
No mocking of internal modules
verify_on_exit! in setup for strict mocking

Test Adapters

Bypass for HTTP endpoint mocking
Swoosh.TestAdapter for email testing
Oban.Testing for background job assertions

Database

Ecto.Adapters.SQL.Sandbox for isolation
Async tests don't share database state
Fixtures/factories used consistently

Valid Patterns (Do NOT Flag)

Mock in unit test, real in integration - Different test levels have different needs
Not mocking database in integration tests - Database is internal
Simple inline test data - Not everything needs factories
Testing private functions via public API - Correct approach

Context-Sensitive Rules

Issue	Flag ONLY IF
Not async	Test actually needs shared state
Missing mock	External call exists AND no mock/bypass
Mock internal	Module being mocked is internal code

Gates (sequence)

Complete in order. Do not emit a finding until the prior step passes for that issue.

Evidence from the file — Open the test module (or helper) and tie the claim to concrete lines.
- Pass when: Each prospective finding includes [FILE:LINE] and a one-line factual description of what is on that line (or an adjacent line you name), not a generic style complaint.
ExUnit false-positive veto — Check this skill’s Valid Patterns and Context-Sensitive Rules for the case.
- Pass when: You can state “not covered by Do NOT Flag / Flag ONLY IF” in one sentence, or you drop the finding.
Cross-protocol verification — Apply review-verification-protocol (e.g. read full function/block, search usages before “unused” claims) to that same finding.
- Pass when: At least one protocol check relevant to the claim type is satisfied and would appear in your rationale if challenged.

Before Submitting Findings

Use [FILE:LINE] ISSUE_TITLE per finding after Gates (sequence) and the linked protocol are satisfied.

安全使用建议

This skill appears coherent for reviewing ExUnit test files: it will read your repository's test modules and helpers to produce line-referenced findings. Before installing, confirm you're comfortable with the agent reading your test files (it does not request credentials or perform installs). If you handle sensitive data in tests, consider running reviews locally or ensuring the agent's model/process does not send repository contents to external services you don't trust. If you want stricter control, you can keep autonomous invocation disabled at the agent level or manually invoke the skill when needed.

功能分析

Type: OpenClaw Skill Name: exunit-code-review Version: 1.2.1 The skill bundle is a legitimate tool designed to guide an AI agent in reviewing Elixir ExUnit test code. It contains comprehensive documentation and checklists in SKILL.md and the references directory (e.g., exunit-patterns.md, mox-boundaries.md) that align perfectly with its stated purpose of enforcing testing best practices. There are no signs of malicious intent, data exfiltration, or prompt injection aimed at compromising the agent or the host system.

能力标签

cryptocan-make-purchases

能力评估

✓ Purpose & Capability

Name and description (ExUnit test review) match the content: the skill is instruction-only, contains reference docs about ExUnit/Mox/Bypass/Oban, and declares no binaries, env vars, or installs that would be unrelated to test review.

✓ Instruction Scope

SKILL.md instructs the agent to open test modules/helpers, produce findings tied to concrete file:line evidence, and perform repository-local checks (read blocks, search usages). Those actions are appropriate and proportionate for a code-review skill; the instructions do not request credentials or external network endpoints.

✓ Install Mechanism

No install spec or code files to run; instruction-only distribution means nothing is written to disk or downloaded during install — lowest-risk delivery mechanism for this purpose.

✓ Credentials

No required environment variables, credentials, or config paths are declared or referenced. The recommended actions (reading test files) are consistent with not needing secrets or external credentials.

✓ Persistence & Privilege

always is false and the skill does not request persistent system changes or modify other skills. It allows normal autonomous invocation (disable-model-invocation: false) which is the platform default and acceptable here.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install exunit-code-review
安装完成后，直接呼叫该 Skill 的名称或使用 /exunit-code-review 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.2.1

- Added "Gates (sequence)" section outlining a precise multi-step review process before submitting findings. - Clarified that each finding must be tied to concrete file evidence, avoid false-positives per "Valid Patterns" and "Context-Sensitive Rules," and pass cross-protocol verification. - Adjusted the instructions for submitting findings to require passing all review gates before reporting issues. - The rest of the documentation, including checklist and philosophy, remains unchanged.

v1.2.0

exunit-code-review 1.2.0 – Enhanced test review coverage and guidance - Improved documentation with detailed mock boundary philosophy and review checklist. - Added quick reference links for exunit patterns, Mox usage, test adapters, and integration strategies. - Clarified proper mocking boundaries: external dependencies only, not internal code. - Expanded checklist for test structure, mocking, adapters, and database handling. - Included context-sensitive flagging rules and clarified when to mock or not. - Specified reporting format and review protocol requirements before submitting findings.

元数据

Slug exunit-code-review

版本 1.2.1

许可证 MIT-0

累计安装 1

当前安装数 1

历史版本数 2

常见问题