Plugin Claude Code
An intelligent development workflow featuring Socratic interviewing, dual-agent debate, expert team coordination, and RAG system design. Built on Claude Code for seamless AI-assisted development.
Plugin Claude Code
A Claude Code workflow made of commands and AI agents for gathering requirements, planning work, writing code, testing, reviewing, debugging, and designing retrieval-based AI systems.
Instructions file CodexOpenCode
Instructions for an8079/take-skills, covering agents.md - agent catalog, agent overview, agent specifications, interviewer and architect.
Instructions file
Project instructions and an index for a Chinese-language Claude Code toolkit. The toolkit contains agents, commands, domain skills, rules, scripts, and a staged workflow from requirements discovery through delivery.
Agent
Agent "team-agents" from an8079/take-skills, covering team agents - 团队协作 agent 定义, agent 总览, agent 详细定义, team-lead and analyst.
Command
An index of coding commands for a staged workflow, from interviewing and planning through implementation, testing, review, and debugging. It also lists commands for project analysis, automation, quality checks, and team work.
Command
A coding assistant persona for a full-stack developer, meaning someone who works across both the visible app and its server-side code. It supports test-driven development, or TDD, where tests are written before implementation and then used to guide changes.
Command
A debugging assistant for analysing error messages, stack traces, logs, inputs, and environments. A stack trace is the list of code calls showing how an error was reached.
Command
A Chinese-language requirements interview assistant that asks detailed questions about a proposed software project. It helps define the problem, boundaries, technical limits, and first usable version.
Command
A planning command that breaks a specification into small tasks, maps their dependencies, identifies risks, sets priorities, and defines acceptance checks.
Command
A code-review command whose reviewer checks whether changes work correctly, stay secure, remain maintainable, perform acceptably, and have suitable tests.
Command
An architecture-design command for planning how software is structured, including its data, interfaces, and major components.
Command
A command that examines an existing project’s files, technologies, entry points, modules, APIs, and data models, then produces an architecture report.
Command
A command that runs a debate between a product manager focused on user and business value and an architect focused on technical design and feasibility.
Command
An automatic task-running command with five stages: expanding the request, planning the work, executing it, checking quality, and validating the result. It repeats the work cycle when the goal has not yet been reached.
Command
A command that puts the agent into a strict boss-like supervision mode, using Chinese-language prompts to judge work quality and progress.
Command
A command that removes unnecessary complexity from AI-generated code while preserving its existing behaviour.
Command
A code-review command that asks four separate reviewers to inspect security, performance, maintainability, and architecture. It combines their findings into one report ordered by severity.
Command
A Socratic requirements interview command that asks questions to turn an unclear idea into a structured specification document.
Command
A command that interviews you about a project, searches the lenny-skills repository, and recommends skills to install.
Command
A command that repeatedly has one agent improve a web page while another agent scores it against measurable criteria. It uses agent-browser to inspect the page.
Command
A command that repeatedly improves a web page by having one agent change it and another agent score it with browser-based checks.
Command
A project-import command that examines an existing codebase and defines which parts an AI development system may change.
Command
A command that assigns GitHub Issue labels by looking for keywords in the issue title and description. It uses GitHub CLI to apply labels for priority, issue type, complexity, and affected component.