开发技能

将功能开发、大规模代码重构和拉取请求审查交给Codex、Claude Code或OpenCode在后台执行，并处理独立检出目录、进度监控及完成或失败通知。

• 将功能开发交给后台代理
• 在独立工作目录中进行大规模重构
• 请编码代理审查拉取请求

按条件筛选GitHub的Issue，交给后台智能体完成修复、测试和PR创建。工作流会检查重复任务，也支持处理审核意见并定期监测新问题。

• 按标签或负责人选择问题
• 为多个问题创建修复PR
• 处理已创建PR的审核意见

通过暂停Node.js执行、检查变量和跟踪执行路径，调查异步任务停滞、测试偶发失败等问题，也可分析内存持续增长和CPU负载的原因。

• 调查异步任务停滞的原因
• 调试偶发失败的测试
• 排查启动和子进程问题

审查REA的分析流程与缓存设计，在保持结果正确性和依据可访问性的同时，改善耗时、CPU与内存使用，以及智能体需要处理的信息量。

• 调查分析流程缓慢的原因
• 审查缓存设计
• 降低CPU和内存负载

Design or change REA investigation tools by defining responsibilities, inputs, results, and CLI or MCP interfaces. Produce a proposal that explains how Evidence records should be interpreted, compares the nearest existing alternative, and outlines verification.

中文简介正在准备中。当前显示英语简介。

• Designing a new REA investigation tool
• Reviewing CLI or MCP inputs and results
• Defining provider responsibilities and Evidence meaning

Investigate shipped applications, binaries, packages, and firmware with REA to explain features, compare versions, or guide reconstruction. Use decompiled code and runtime observations where needed, keeping observed facts, inferred behavior, and unresolved questions clearly separated.

中文简介正在准备中。当前显示英语简介。

• Explaining features in a shipped application
• Comparing behavior across application versions
• Inspecting binaries and packaged resources

Examine superpowers session records to document repeated work, departures from plans, delays, and token usage. Produce an evidence-based report with file and line references, and prepare bug-report material when requested.

中文简介正在准备中。当前显示英语简介。

• Investigating repeated work or departures from a plan
• Checking session duration and token usage
• Tracing whether a skill was invoked

Investigate bugs, failing tests, and unexpected behavior using reproduction steps, logs, and recent changes. Follow a structured process to identify the root cause, test a specific explanation, and verify the resulting fix.

中文简介正在准备中。当前显示英语简介。

• Investigating failing tests
• Tracing the cause of a production bug
• Finding where a build or integration breaks

Requests a code review from a separate agent after implementation or before merging. It provides the change summary, requirements, and commit range, then guides follow-up work according to the severity of the findings.

中文简介正在准备中。当前显示英语简介。

• Reviewing a major feature after implementation
• Checking changes before merging
• Getting another perspective on a complex bug fix

Create, revise, and validate agent skills through tests that compare behavior with and without the instructions. It helps identify unclear guidance, choose an appropriate document structure, and refine wording before deployment.

中文简介正在准备中。当前显示英语简介。

• Creating a reusable agent skill
• Revising instructions agents misinterpret
• Testing a skill before deployment

Carry out an existing implementation plan in one session, task by task. Verify each step with tests, record progress and decisions in a file, and finish with a review of all changes on the branch.

中文简介正在准备中。当前显示英语简介。

• Implementing a prepared development plan
• Executing tasks without implementer subagents
• Resuming work from a progress record

将实施计划中的任务逐一交给不同的代理，检查是否符合规格要求以及代码质量，并将进度记录到文件中，最后审查整个分支的全部变更。

• 执行任务基本独立的实施计划
• 逐项检查任务是否符合规格要求
• 根据进度记录恢复工作

Assign independent bugs or test failures to separate agents that work at the same time. Define each agent’s scope and expected report, then check for conflicting changes and run the full test suite before integrating fixes.

中文简介正在准备中。当前显示英语简介。

• Investigating test failures with unrelated causes
• Fixing independent problems in separate subsystems
• Defining focused tasks for multiple agents

Turn specifications or requirements into an implementation plan before coding begins. Break work into testable tasks with exact files, function definitions, test assertions, and verification commands, so another developer or agent can follow the plan.

中文简介正在准备中。当前显示英语简介。

• Planning a feature before writing code
• Breaking requirements into testable tasks
• Handing implementation work to another developer or agent

Evaluate code review feedback against the actual code, supported environments, and existing behavior. Clarify uncertain requests, respond with technical evidence, and implement valid changes one at a time with tests for each fix.

中文简介正在准备中。当前显示英语简介。

• Clarifying ambiguous code review comments
• Checking whether suggested changes preserve compatibility
• Assessing whether a requested feature is actually used

Prepare a separate workspace for feature development using platform tools or git worktree, which gives a branch its own working directory. It checks existing isolation, prepares dependencies, and runs tests before implementation begins.

中文简介正在准备中。当前显示英语简介。

• Starting feature work in a separate workspace
• Preparing a workspace before executing an implementation plan
• Checking whether a workspace is already isolated

Finish completed development work by checking the full test suite and presenting integration choices: merge locally, create a Pull Request, or keep the branch. It carries out the selected path and handles workspace cleanup accordingly.

中文简介正在准备中。当前显示英语简介。

• Merging a completed feature locally
• Pushing a branch and creating a Pull Request
• Keeping completed work for later integration

Turn an existing discussion and codebase knowledge into a specification, then publish it to the project's issue tracker. Organize the problem, user needs, implementation decisions, testing approach, and work excluded from scope.

中文简介正在准备中。当前显示英语简介。

• Turning a feature discussion into a specification
• Publishing agreed requirements to the project issue tracker
• Documenting implementation and testing decisions

Review a codebase for structural problems that make understanding and testing difficult. Compare refactoring candidates in an HTML report with before-and-after diagrams, then explore the chosen option through questions about constraints, design, and tests.

中文简介正在准备中。当前显示英语简介。

• Review architecture in frequently changed code
• Compare refactoring options with before-and-after diagrams
• Explore module designs that are easier to test

Turn a course plan into numbered exercise folders with problem, solution, and explanation variants. Create starter readme files, check naming and links with the project’s linter, and fix structural errors before committing.

中文简介正在准备中。当前显示英语简介。

• Set up exercise folders for a new course section
• Create problem, solution, and explanation stubs
• Renumber exercises and check their structure

Write a pull request description that helps reviewers understand a code change. Use a compact diagram or diff, show before-and-after evidence, and explain how reversible the change is and what areas it may affect.

中文简介正在准备中。当前显示英语简介。

• Writing a pull request description
• Explaining a refactor with a file tree or diff
• Presenting before-and-after evidence for a UI change

Review code changes from a chosen Git reference against two separate criteria: documented coding standards and the original requirements. Independent reviewers check each criterion, then present findings separately so convention problems and requirement gaps remain visible.

中文简介正在准备中。当前显示英语简介。

• Reviewing changes on a branch or pull request
• Checking changes since a chosen commit or tag
• Checking compliance with documented coding standards

Review a coding session and identify improvements to the agent’s working environment. Examine file navigation, automated checks, instructions, tool use, and information access, then present proposed changes in order of severity.

中文简介正在准备中。当前显示英语简介。

• Reviewing a coding session that took too long
• Finding automated checks for recurring mistakes
• Simplifying agent instructions

Implement work described in a specification or tickets, checking types and running tests along the way. Review the completed changes with the code-review skill, then commit the implementation to the current branch.

中文简介正在准备中。当前显示英语简介。

• Implementing a feature from a specification
• Implementing a referenced issue tracker ticket
• Checking changes with types and tests