测试与安全技能

在发布前验证OpenClaw插件的安装、卸载、配置修复、启动、SDK兼容性及分发包行为，记录测试结果、失败情况与尚未验证的项目。

• 发布前检查内置插件的安装与卸载
• 在干净环境中验证测试版
• 检查配置修复与gateway启动

Process OpenClaw GitHub Secret Scanning alerts, which flag exposed credentials. Inspect affected content, redact secrets, replace comments when needed, notify authors, and resolve alerts using a dedicated script under maintainer permissions.

中文简介正在准备中。当前显示英语简介。

• Investigating locations flagged for exposed credentials
• Redacting secrets in issue or pull request descriptions
• Deleting and recreating comments containing secrets

Assess OpenClaw security reports against released versions and documented trust boundaries, which define trusted users and components. Gather evidence to recommend closing, keeping open, or narrowing a report, and draft a maintainer response.

中文简介正在准备中。当前显示英语简介。

• Assessing an OpenClaw security report
• Checking whether released versions contain a reported bug
• Comparing a report with existing advisories

Investigate memory growth and out-of-memory failures in OpenClaw tests. Compare memory measurements and heap snapshots, records of objects held in memory, to distinguish cleanup leaks from retained test modules and verify targeted fixes.

中文简介正在准备中。当前显示英语简介。

• Investigating Vitest out-of-memory failures
• Comparing heap snapshots from the same process
• Distinguishing cleanup leaks from retained test modules

Test OpenClaw on Telegram's Test Server through a real user account. Record replies, edits, deletions, reactions, and typing to verify user-visible behavior, with actual Telegram client screenshots required when checking visual results.

中文简介正在准备中。当前显示英语简介。

• Verifying replies and commands in direct messages
• Testing group mentions, topics, and reactions
• Tracking message edits, deletions, and typing

Measure and improve OpenClaw test and plugin-suite performance by investigating slow imports, repeated startup work, and memory use. Compare results before and after changes while preserving test coverage and verifying the affected tests.

中文简介正在准备中。当前显示英语简介。

• Finding slow tests and expensive module imports
• Reducing plugin-suite runtime and memory use
• Investigating growing memory use during tests

Verify two-way Discord messaging in a macOS virtual machine running in Parallels. Check installation and gateway health, confirm messages reach the host, and prove that the guest can read a new host reply.

中文简介正在准备中。当前显示英语简介。

• Checking Discord messaging after a fresh macOS installation
• Verifying two-way delivery after an upgrade
• Investigating intermittent Discord roundtrip failures

Test Slack authentication, message creation and cleanup, and OpenClaw message handling. Collect evidence from stored messages and test runs, while distinguishing what API checks prove from interactions that require observation in the actual Slack client.

中文简介正在准备中。当前显示英语简介。

• Check Slack authentication and channel access
• Test message creation, editing, and deletion
• Verify OpenClaw message handling and replies

Review new and existing software tests for the behavior and regressions they actually protect. Identify duplicate or implementation-dependent checks, document evidence before removal, and simplify code used only by tests while retaining meaningful coverage.

中文简介正在准备中。当前显示英语简介。

• Checking whether a new test adds useful coverage
• Auditing duplicate or implementation-dependent tests
• Reviewing code exposed only for testing

Choose suitable tests and checks for OpenClaw changes, then investigate failures using the relevant evidence. It guides local validation and directs environment-dependent or release testing to the appropriate procedures, keeping checks tied to the changed behavior.

中文简介正在准备中。当前显示英语简介。

• Choosing checks for source code changes
• Reproducing a bug and verifying its repair
• Investigating failed CI jobs

Create and update OpenClaw end-to-end tests in Docker, checking complete workflows. Use fake servers to verify requests and real providers to check service behavior, including image understanding and the final text users receive.

中文简介正在准备中。当前显示英语简介。

• Adding or updating Docker end-to-end tests
• Checking payloads sent to dependencies
• Verifying behavior with real model providers

为人工测试OpenClaw候选版本准备隔离的OCM副本或经批准的现有Gateway环境，指导测试、整理反馈，并协助在公开报告前检查内容。

• 使用Gateway隔离副本进行测试
• 由人工逐一测试操作界面
• 公开前检查验证报告

通过界面状态对比图库和真实浏览器操作测试，检查OpenClaw Control UI的界面改动，保存用户评论以及改动前后经检查的截图。

• 对比加载、空白和错误状态
• 收集对界面改动的意见
• 通过模拟Gateway验证聊天操作

持续开展OpenClaw整体质量验证，从复现缺陷、修复根因到独立审查和确认合并，将真实环境及压力测试的证据和进度记录在可续接的报告中。

• 调查OpenClaw多个领域的缺陷
• 通过回归测试和独立审查验证修复
• 验证实际连接、界面及分发包

使用Discord测试机器人核查连接信息和读取权限，并通过证据验证消息、文件、讨论串操作及OpenClaw响应，同时检查结果日志和测试对象的清理情况。

• 核查机器人身份和历史记录读取权限
• 验证消息、文件和讨论串操作
• 验证OpenClaw回复、流式发送及输入提示

Audit and refresh OpenClaw maturity scorecards using feature definitions, recorded scores, and QA evidence. Review quality, workflow completeness, and long-term support status, then validate source data and regenerate public documentation.

中文简介正在准备中。当前显示英语简介。

• Review maturity scores for a feature area
• Assess whether expected user workflows are complete
• Update scorecard documentation from QA evidence

Review and update OpenClaw security advisories on GitHub, check linked private forks and publication requirements, and verify the published result. The workflow includes approval before publication and checks for common update errors.

中文简介正在准备中。当前显示英语简介。

• Reviewing an advisory and its linked private fork
• Updating vulnerability details and severity
• Checking publication requirements before approval

Run installation, integration, and end-to-end tests that follow real user actions in remote environments or Docker. Choose an appropriate isolated environment, inspect failures, record execution details, and clean up resources after testing.

中文简介正在准备中。当前显示英语简介。

• Testing package installation in a clean environment
• Verifying an integration with a real service
• Checking behavior across macOS, Windows, and WSL2

使用模拟环境或真实服务运行OpenClaw质量验证场景，调查并修复失败原因、检查结果报告，还可比较多个模型的表达方式和角色表现。

• 在模拟环境或真实服务中运行质量验证
• 调查失败场景并重新验证
• 比较多个模型的角色表现

Test the delivery order of Telegram drafts and final messages in OpenClaw's QA Lab. Run the same test scenario through QA Channel or the real Telegram plugin to collect evidence about streaming and message delivery.

中文简介正在准备中。当前显示英语简介。

• Checking the delivery order of Telegram drafts and final messages
• Running message flow tests through QA Channel
• Testing the same scenario with the real Telegram plugin

Verify OpenClaw installation, onboarding, gateway operation, and upgrades in Parallels virtual machines running macOS, Windows, or Linux. The skill guides baseline restoration, bounded test runs, log interpretation, and focused retries when checks fail.

中文简介正在准备中。当前显示英语简介。

• Checking fresh OpenClaw installs across three operating systems
• Verifying upgrades through the installed OpenClaw updater
• Restoring existing VM snapshots for repeatable checks

Assess the security of a machine running OpenClaw through read-only checks. Review remote access, firewall settings, updates, backups, disk encryption, and gateway exposure, then propose staged improvements with evidence and rollback guidance.

中文简介正在准备中。当前显示英语简介。

• Checking the security of an OpenClaw machine
• Reviewing a server's public network exposure
• Checking backups and disk encryption

Audit REA’s automated checks and test suites to find missed failures, outdated test data, and unnecessary work. Use reproducible experiments and timing measurements to assess fixes or efficiency changes while preserving checks that catch real defects.

中文简介正在准备中。当前显示英语简介。

• Investigating CI checks that pass despite defects
• Checking whether test data matches current behavior
• Measuring unnecessary work in slow test suites

Guides feature development and bug fixes by writing tests before implementation. It confirms each test fails for the expected reason, adds only enough code to pass, and checks the full test suite before completion.

中文简介正在准备中。当前显示英语简介。

• Implementing a new feature from tests
• Reproducing a bug in a test before fixing it
• Refactoring while keeping tests passing