Claude Code 是什么
详解这款智能体式 CLI/IDE 工具——会话循环、安装方式，以及各项功能的所在位置。
Claude Code Cheat Sheet
One printable page of Claude Code flags, commands, shortcuts, permission modes and file locations — every line checked against the official docs on 2026-10-13.
claude CLI：真正会改变结果的标志
对 Claude Code 命令行的一次有主见的巡览——无人值守运行的五标志约定、已记录系统提示词的陷阱，以及 --bare、--safe-mode 和 --restricted 的区别。
CLAUDE.md 与记忆文件
收益最高的 Claude Code 定制：持久化的项目记忆。
AGENTS.md 与跨工具互操作
每个编码代理都会读取的开放标准——以及如何让它与你的 CLAUDE.md 保持同步。
规划模式
在任何编辑执行之前先调查并提出方案——处理大型改动最安全的方式，以及它并不强制执行的限制。
IDE 集成
VS Code 与 JetBrains 扩展、内联 diff，以及共享设置。
settings.json：配置系统
文件层级与优先级、关键小节，以及项目配置与个人配置。
权限与权限模式
allow/ask/deny 规则与模式——在不牺牲安全的前提下减少确认提示。
Agent SDK Permissions: the Six-Step Order
Every permission bug in a headless agent is an order bug. The six steps the Agent SDK evaluates — which ones can only deny, which can only allow, and the five ways allowedTools silently does nothing.
Sandbox Credential Masking: keep tokens working, keep them secret
Mask environment variables, JWTs, AWS SigV4 pairs, and credential files so sandboxed commands authenticate without ever seeing the real secret. Config, gotchas, and OS matrix.
自定义斜杠命令
把一个可重复的工作流变成一个词。
The /design Command: Artboards Before Code
Claude Code's /design skill drafts editable UI artboards before writing a line — how it works, when it wins, and the gotchas nobody documents.
输出风格
改变 Claude Code 的语气、详略程度和讲解行为。
状态栏定制
用自定义脚本展示模型、当前目录、上下文预算和 git 状态。
使用 Spec Kit 进行规格驱动开发
别再用氛围编程硬上大功能。把可执行的规格当作唯一事实来源，然后让 Claude Code 据此构建——借助 GitHub 的 Spec Kit。
上下文管理
/compact 与 /clear、自动压缩，以及随着会话增长保持高效。
Context Budget Audit: /skill-doctor, Output Caps and the Six Numbers That Fill Your Window
A 20-minute audit for a Claude Code project that has grown slow, forgetful, or expensive. Measure with /context and /usage, prune the skill listing with /skill-doctor and skillOverrides, cap tool output with bashOutputMaxChars and taskOutputMaxChars, keep CLAUDE.md out of subagents with omitClaudeMd, and set the auto-compact window on purpose. Every number cited to the September 2026 Claude Code docs.
检查点与回溯
撤销 Claude 的编辑、回滚整个会话，并用 /rewind 压缩上下文——为大胆运行提供的安全网。
Hooks：确定性自动化
在生命周期事件上运行命令，以强制执行格式化、关卡检查和安全规则。
技能（Skills）：按需调用的专长
SKILL.md、渐进式披露，以及技能 vs 命令 vs 子代理 vs MCP。
子智能体与并行智能体
把工作委派给上下文隔离、工具受限的智能体——以及何时不该这么做。
Subagent Fleet Limits: Concurrency Caps & Nested Depth
On July 21, 2026 Claude Code capped concurrent subagents at 20 and killed nested spawning. Three days later nesting came back at depth 3. Here are the four env vars, the error you'll hit, why the caps exist, and how to design fleets that survive them.
动态工作流与 ultracode
从 Claude 编写的脚本中编排数十到数百个子代理——用于代码库审计、大规模迁移和交叉核验式研究。
Cross-Session Messaging
Make two Claude Code sessions talk to each other — the native SendMessage / ListAgents tools shipped in Claude Code 2.1.224, the @mention syntax, the trust boundary, and when to use messaging vs. subagents vs. worktrees.
Projects: The Coordinator & Parallel Threads
On September 17, 2026 Claude Code Projects became an orchestration layer: a coordinator that scopes and delegates, and threads that are full cloud sessions, each on its own branch and repo copy. Here is the architecture, how it differs from subagents and worktrees, what it costs against your usage limits, and the eligibility rule that explains why you may not have it yet.
Claude Code 中的 MCP 服务器
通过模型上下文协议连接数据库、GitHub、浏览器等等。
后台任务、/loop 与调度
长时间运行的命令、周期性循环，以及定时云智能体。
Auto-Commit Background Agents: Draft-PR Handoff
On July 1, 2026 Claude Code v2.1.198 let background agents commit, push, and open a draft PR on their own. Five weeks later v2.1.221 tuned the default. Here is what the flow actually does, the notification hook it fires, why the community wanted an off switch, and the propose-only pattern that still works.
插件与市场
把技能、智能体、命令、钩子和 MCP 打包成一个可安装的单元。
Mods: In-Process Extensions
Mods are TypeScript middleware that runs inside Claude Code — rewrite tool calls, prompts, turns, and the interface itself.
Plugin Evals：用 claude plugin eval 测试你的插件
证明 Claude Code 插件确实有用：编写评测用例，用六种 grader 打分，读取有/无插件的差值，模拟 MCP 服务器，并以分数作为 CI 的关卡。
Git Worktree 与并行工作流
在同一个仓库上运行多个相互隔离的 Claude 会话，互不冲突。
Parallel Agents at Merge Time: Intent Conflicts and the CI Wall (2026)
Worktrees stop agents from overwriting each other's files. They don't stop two plans that can't both be true from merging cleanly. The coordination layer above Git (Foremerge's protocol, decoded) plus the CI surgery Linear did when agents started adding 2,000 tests a week — with every number.
无头模式与 Agent SDK
在 CI 中运行 Claude Code，并在同一套底座上构建生产级智能体。
Cloud Sessions: the VM Claude Code runs in when it isn't your laptop
Practitioner guide to Claude Code cloud sessions and cloud environments — exactly what carries over from your machine and what silently doesn't, the four network access levels and the four paths that bypass them, network secrets, setup scripts vs SessionStart hooks, the ~5-minute cache rule, resource and time limits, and the --cloud/--teleport round trip.
Self-Hosted Environments: run cloud sessions on your own boxes
Practitioner guide to the claude self-hosted-runner beta (shipped Aug 7, 2026 in v2.1.224). Run Claude Code web, mobile and desktop cloud sessions on infrastructure you control — egress-only, no inbound holes, one-user-at-a-time isolation. Architecture, quickstart, runner lifecycle, git creds, autoscaling, and the six gotchas that catch every first fleet.
The MCP Token Tax
Why connecting MCP servers can burn 50k–150k tokens before your first request — and the three fixes: Tool Search, code execution, and knowing when a CLI call wins.
Accessibility
Screen reader mode, vim ergonomics, mouse and scroll controls, and high-contrast themes in Claude Code.
Troubleshooting Claude Code
Symptom-to-fix for hangs, high memory, broken search, compaction thrashing, and garbled terminals.
Feature Flags: Why Your Claude Code Isn't Mine
Claude Code switches features on with flags it fetches from Anthropic. Privacy env vars, Bedrock, gateways and fresh installs silently turn that fetch off — here's exactly what you lose and how to pin behavior instead.