编码规范、框架、重构、代码审查（1.8万 个）
Modify an existing file by replacing an exact text match with new text, optionally changing every match. This focused editing method supports Windows and Unix line endings and is suited to small, targeted updates such as renaming a function or variable.
中文简介正在准备中。当前显示英语简介。
Create Mermaid diagrams for workflows, system architecture, database models, schedules, and other common patterns. Choose a visual theme, render as SVG, PNG, or plain text ASCII, and generate a related set of diagrams in one batch.
中文简介正在准备中。当前显示英语简介。
Investigate bugs, failing tests, and unexpected behavior by reproducing the issue, checking evidence, and tracing its cause before proposing a fix. The skill guides hypothesis testing and verification so changes target the underlying problem.
中文简介正在准备中。当前显示英语简介。
Turn an early idea into an agreed design before implementation. The skill reviews project context, asks focused questions, compares possible approaches, and records the approved design so work can move into planning.
中文简介正在准备中。当前显示英语简介。
Delete a specified file or an empty directory. The skill rejects non-empty directories, so removing a directory tree requires a separate shell command. Use it for targeted cleanup when you know the exact file or empty folder to remove.
中文简介正在准备中。当前显示英语简介。
Create interactive, multi-component web artifacts with React, TypeScript, Tailwind CSS, and shadcn/ui. The workflow initializes a project, supports state and routing, and bundles the finished work into one shareable HTML file for Claude conversations.
中文简介正在准备中。当前显示英语简介。
Turn a specification or requirements for a multi-step development task into a detailed implementation plan. Break work into small actions and identify files, code, tests, documentation, and commands needed for each step.
中文简介正在准备中。当前显示英语简介。
Use this skill to guide feature work, bug fixes, and refactoring with test-driven development. It has the agent write a focused test, confirm that it fails for the expected reason, add the smallest implementation, and refactor while keeping tests green.
中文简介正在准备中。当前显示英语简介。
Evaluate code review feedback against the codebase before changing anything. Clarify unclear requests, check technical correctness and possible regressions, then implement valid fixes individually or explain a reasoned pushback when a suggestion does not fit.
中文简介正在准备中。当前显示英语简介。
Add and manage MCP servers in OpenAkita, from local npm or Python packages to remote HTTP and SSE services. Configure how each server starts or connects, reload settings, check connection status, and troubleshoot common setup problems.
中文简介正在准备中。当前显示英语简介。
Build MCP servers that let AI models use external APIs and services through clearly described tools. The guide covers planning, implementation, review, testing, and evaluations in Python with FastMCP or TypeScript with the MCP SDK, then registering the finished server in OpenAkita.
中文简介正在准备中。当前显示英语简介。
Review local code changes or a remote Pull Request for correctness, security, maintainability, readability, efficiency, edge cases, and test coverage. It organizes feedback by severity and provides a clear recommendation before changes are merged.
中文简介正在准备中。当前显示英语简介。
Turn Git commits into clear release notes for customers. The skill reviews changes over a chosen period or between versions, groups features and fixes, and filters internal work to create structured updates for changelogs, app stores, or product announcements.
中文简介正在准备中。当前显示英语简介。
Design and implement working frontend interfaces such as websites, landing pages, dashboards, and React components. The skill helps shape a visual direction around the interface’s purpose and constraints, then refines typography, color, layout, motion, and details in code.
中文简介正在准备中。当前显示英语简介。
Manage GitHub repositories, issues, pull requests, and branches, and search code or check CI and deployment status through Rube MCP. It supports common repository workflows, with safeguards such as checking pull request status before merging.
中文简介正在准备中。当前显示英语简介。
Describe an app in natural language to generate and deploy it through a conversational workflow. The skill covers web apps, WeChat mini programs, games, AI tools, SaaS products, and dashboards, with follow-up requests for changes such as adding user login.
中文简介正在准备中。当前显示英语简介。
Run ClawSweeper reviews locally for an existing issue or pull request, or for a committed range before submission. The skill guides the review command and artifact review while keeping GitHub unchanged; the range mode still calls the configured Codex model service.
中文简介正在准备中。当前显示英语简介。
Use this skill when designing server-side applications with Node.js, Express, or Next.js API routes. It provides patterns for API structure, database queries, caching, authentication, errors, rate limits, background jobs, and logging to help organize backend code.
中文简介正在准备中。当前显示英语简介。
Use this skill to guide TypeScript, JavaScript, React, and Node.js work with shared coding conventions. It covers readable naming, types, error handling, API design, project structure, performance, testing, and common code smells to review or refactor.
中文简介正在准备中。当前显示英语简介。
Use this pattern to gather context for multi-agent work when the needed files and project terms are not known in advance. Start with a broad search, assess relevance and missing information, then refine the query for up to three cycles.
中文简介正在准备中。当前显示英语简介。
Use this PostgreSQL reference when designing schemas, choosing indexes, or investigating slow queries. It also shows patterns for row-level security, pagination, queue processing, connection settings, and checks for common database issues.
中文简介正在准备中。当前显示英语简介。
Use this skill when building React or Next.js interfaces. It provides patterns for composing components, managing state, fetching data, validating forms, improving rendering performance, handling errors and animation, and supporting keyboard navigation and focus management.
中文简介正在准备中。当前显示英语简介。
Check code changes through build, type, lint, test, security, and diff review steps. The skill organizes findings into a report that shows which checks passed, what issues remain, and whether the work is ready for a pull request.
中文简介正在准备中。当前显示英语简介。
Help multi-agent coding workflows find useful repository context without passing every file at once. Search broadly, assess file relevance and missing information, then refine the query for up to three cycles to produce a focused set of context for a sub-agent.
中文简介正在准备中。当前显示英语简介。