Claude Blog 采集 (2026-06-01)¶

共采集 5 篇文章

📋 文章索引¶

How and when to use subagents in Claude Code - Apr 07, 2026 (评分: 9.5)
Claude Enterprise, now available self-serve - Feb 12, 2026 (评分: 9.0)
Introducing Agent Skills - Oct 16, 2025 (评分: 9.0)
Structured outputs on the Claude Developer Platform - Nov 14, 2025 (评分: 9.0)
Introducing the Claude Team plan and iOS app - May 01, 2024 (评分: 9.0)

How and when to use subagents in Claude Code¶

来源: Claude Blog 发布日期: Apr 07, 2026 采集时间: 2026-06-01 价值评分: 9.5/10 正文字数: ~20131 字符

摘要¶

When to delegate research, parallelize tasks, or get a fresh review with Claude Code subagents—and when to stick with the main session.

正文内容¶

A practical guide to Claude Code subagents: when they help, how to direct them, and the signals that tell you delegation is worth it.

Claude Code handles complex, multi-step projects well, but long sessions accumulate weight. Every file read, every tangent explored, every half-finished thought stays in the context window, slowing responses and driving up token costs. Consider building a new feature in a large TypeScript monorepo. The main work is the implementation, but side tasks keep appearing: trace how an existing service handles auth, find the shared util for date formatting, check whether the design system already has a component close to what you need. None of these need the full project context, and running them inside the main session adds noise. What if you could run them in parallel? Enter subagents . A subagent is an isolated Claude instance with its own context window. It takes a task, does the work, and returns only the result. Think of subagents as the browser tabs of a Claude Code session: a place to chase a tangent without losing the main thread. In this article, we discuss when it makes sense to use subagents, how to invoke them, and when the overhead isn't worth it. What is a subagent? Subagents are self-contained agents that operate with their own context windows. When Claude spawns a subagent, that assistant works independently to read files, explore code, or make changes. When it completes its task, the subagent returns only the relevant results to the main conversation. Each subagent starts fresh, unburdened by the history of the conversation or invoked skills. Multiple subagents can run in parallel, and each can have different permissions: a research subagent might have read-only access, while an implementation subagent gets full editing capabilities. Claude Code includes several built-in subagent types, including: General-purpose agents for complex multi-step tasks Plan agents that research codebases before presenting implementation strategies Explore agents optimized for fast, read-only code search Claude Code often spawns subagents on its own to handle assigned tasks. It's also possible to direct that behavior explicitly and to define reusable specialists that Claude delegates to automatically. Knowing when to reach for subagents is what makes the feature useful. When should you use subagents? Certain categories of work benefit clearly from subagent delegation. Learning to recognize them makes the feature far more effective. Research-heavy tasks When understanding how something works is a prerequisite to changing it, a subagent can explore the codebase and return a summary rather than dumping dozens of files into the conversation. The signal: Gathering context requires reading dozens of files. The benefit: The main conversation stays clean, and synthesized findings arrive instead of raw content. Multiple independent tasks When fixing errors across several files, updating patterns in multiple components, or making changes that don't depend on each other, parallel subagents complete the task faster. The signal: Sub-tasks have no dependencies between them. The benefit: Three subagents working simultaneously generally finish the task in less time. Fresh perspective needed When an unbiased review of an implementation is the goal, a subagent provides a clean slate because it doesn't inherit the assumptions, context, or blind spots from the primary conversation. The signal: Verification is needed without conversation history influencing the analysis. The benefit: Cleaner, more objective feedback. Pro-tip: The /clear command also resets context and conversation history, providing a similarly unbiased slate, but at the cost of losing that history entirely. A subagent achieves the same fresh perspective while the main conversation stays intact. Verification before committing Before finalizing changes, an independent subagent can verify the implementation isn't overfitting to tests or missing edge cases. The signal: A second opinion is warranted before committing code. The benefit: Catches issues that familiarity with the code might obscure. Pipeline workflows When a task has distinct phases (i.e., design, then implement, then test), each stage benefits from focused attention. The signal: Sequential stages with clear handoffs. The benefit: Each subagent concentrates on its phase, without context from other stages creating noise. ‍ Pro-tip: When a task requires exploring ten or more files, or involves three or more independent pieces of work, that's a strong signal to direct Claude toward subagents. How to direct subagent usage Several methods exist for invoking subagents, ranging from simple conversation to automated workflows. The right starting point depends on the workflow, and sophistication can be layered on as patterns emerge. Conversational invocation The most flexible approach is simply asking Claude to use subagents in conversation. This works across all Claude Code interfaces: terminal, VS Code, JetBrains, the web, and desktop applications. Natural language patterns that reliably invoke subagents include: "Use a subagent to explore how authentication works in this codebase" "Have a separate agent review this code for security issues" "Research this in parallel. Check the API routes, database models, and frontend components simultaneously" "Spin up subagents to fix these TypeScript errors across the different packages" Being explicit matters. Specify the scope, request parallel execution when tasks are independent, and describe the desired output. Here's an effective prompt structure: Use subagents to explore this codebase in parallel: 1. Find all API endpoints and summarize their purposes 2. Identify the database schema and relationships 3. Map out the authentication flow Return a summary of each, not the full file contents. This prompt works because it clearly defines three independent tasks, explicitly requests parallel execution, and specifies the output format. Claude understands the intent and spawns appropriate subagents. Tips for effective conversational invocation include: Scope tasks clearly. "Explore how payments work" beats "explore everything." Request parallelization explicitly. Say "these can run in parallel" or "work on all three simultaneously." Specify what should be returned. Summaries, specific findings, or recommendations. Naming the output format helps Claude deliver it. Ask for fresh context when unbiased analysis matters. "Use a subagent that does not see our previous discussion" ensures clean evaluation. Pro-tip: When a subagent is taking a while, Ctrl+B sends it to the background. The conversation can continue while it runs, and results surface automatically when it finishes. The /tasks command shows anything running in the background. Custom subagents When the same kind of subagent keeps getting requested (a security reviewer, a test writer, a docs proofreader), it can be defined once as a custom subagent. Claude then delegates to it automatically whenever a task matches its description, no prompting required. Custom subagents live as markdown files in .claude/agents/ (project-level, shared with the team) or ~/.claude/agents/ (user-level, available across all projects). Each one gets its own system prompt, tool permissions, and optionally its own model. The easiest way to create one is the /agents command, which walks through setup interactively and can generate a first draft from a description. The file can also be written by hand, for example: --- name: security-reviewer description: Reviews code changes for security vulnerabilities, injection risks, auth issues, and sensitive data exposure. Use proactively before commits touching auth, payments, or user data. tools: Read, Grep, Glob model: sonnet --- You are a security-focused code reviewer. Analyze the provided changes for: - SQL injection, XSS, and command injection risks - Authentication and authorization gaps - Sensitive data in logs, errors, or responses - Insecure dependencies or configurations Return a prioritized list of findings with file:line references and a recommended fix for each. Be critical. If you find nothing, say so explicitly rather than inventing issues. With this in place, Claude routes matching work to the subagent automatically. It can also be invoked by name: "Have the security-reviewer look at the staged changes." Custom subagents work best when: A specialist should be available for Claude to delegate to automatically when a task matches The work benefits from a tightly scoped system prompt and restricted tools The configuration should be shared across a team or reused across projects Pro-tip: The description field is what Claude uses to decide when to delegate. Be specific about the trigger conditions, not just the capability. "Reviews code for security issues before commits" routes better than "security expert." For the full configuration reference, including permission modes and how project and user subagents interact, see our Claude Code subagents docs. CLAUDE.md instructions Custom subagents define who the specialists are. CLAUDE.md files define the rules for when Claude should reach for them. If every code review should go through a read-only subagent, or every architecture question should trigger a research pass first, CLAUDE.md is where that policy lives. Claude reads it at the start of every conversation, so the behavior stays consistent across sessions and teammates without anyone needing to remember to ask. CLAUDE.md is a good fit for subagent instructions when: Code reviews should always use read-only subagents The project has specific research patterns Claude should follow Consistent behavior is needed across team members and sessions Here’s an example of a simple CLAUDE.md file that triggers a subagent given specific conditions: ## Code review standards When asked to review code, ALWAYS use a subagent with READ-ONLY access (Glob, Grep, Read only). The review should ALWAYS check for: - Security vulnerabilities - Performance issues - Adherence to project patterns in /docs/architecture.md Return findings as a prioritized list with file:line references. With the above CLAUDE.md file, every code review request automatically uses the defined pattern, eliminating the need to specify it each time. For more on CLAUDE.md files, see Customizing Claude Code for your codebase: setting up a CLAUDE.md file and our Claude Code CLAUDE.md file docs . Skills For complex multi-step workflows that run repeatedly, skills provide a reusable interface. Define a skill once in .claude/skills/, then invoke it with /skill-name or let Claude load it automatically when a task matches its description. Skills differ from CLAUDE.md files in scope. CLAUDE.md files are always loaded and shapes every interaction. A skill is loaded on demand, either because it was invoked explicitly or because Claude matched the current task to the skill's description field. That makes skills the right place for workflows that should be available but not applied to every prompt. Skills fit well when: Certain actions get run regularly Different team members need access to the same complex operation Standardizing how certain tasks are performed across the team matters Here’s an example of a deep-review skill for comprehensive code review: # .claude/skills/deep-review/SKILL.md --- name: deep-review description: Comprehensive code review that checks security, performance, and style in parallel. Use when reviewing staged changes before a commit or PR. --- Run three parallel subagent reviews on the staged changes: 1. Security review - check for vulnerabilities, injection risks, authentication issues, and sensitive data exposure 2. Performance review - check for N+1 queries, unnecessary iterations, memory leaks, and blocking operations 3. Style review - check for consistency with project patterns documented in /docs/style-guide.md Synthesize findings into a single summary with priority-ranked issues. Each issue should include the file, line number, and recommended fix. In the code snippet above, /deep-review triggers a three-part subagent analysis on demand. Because the description mentions reviewing staged changes before commits, Claude can also reach for this skill automatically when that context comes up. A skill is a directory, not a single file. Alongside SKILL.md, it can hold templates Claude fills in, example outputs showing the expected format, or scripts Claude executes as part of the workflow. The legacy .claude/commands/ format was a single flat file, so everything had to live in the prompt itself. For more on using skills with Claude Code, see our Claude Code skills docs. Hooks Hooks are user-defined shell commands, HTTP endpoints, or LLM prompts that execute automatically at specific points in Claude Code's lifecycle. Hooks can automate subagent workflows based on events. Hooks trigger on specific actions and run subagent tasks without manual invocation. Hooks are the right tool when: Every commit should be reviewed automatically before it's created Security checks should run without anyone remembering to ask CI-like quality gates belong in the local development process Here is an example of a Stop hook that blocks Claude from ending its turn until a test is passed: { "hooks" : { "Stop" : [ { "hooks" : [ { "type" : "command" , "command" : "\"$CLAUDE_PROJECT_DIR\"/.claude/hooks/check-tests.sh" } ] } ] } } And the script at .claude/hooks/check-tests.sh : #!/bin/bash INPUT=$(cat) STOP_HOOK_ACTIVE=$( echo " $INPUT " | jq -r '.stop_hook_active // false' ) # Don't loop forever — if we already blocked once this turn, let it through if [ " $STOP_HOOK_ACTIVE " = "true" ]; then exit 0 fi if ! npm test --silent > /dev/null 2>&1; then jq -n '{ decision: "block", reason: "Tests are failing. Run npm test to see the failures and fix them before finishing." }' exit 0 fi exit 0 When Claude finishes its turn, the Stop event fires. The script runs the test suite—if tests fail, it returns JSON with decision: "block" and a reason . Claude Code reads that, doesn't let Claude stop, and feeds the reason back into the conversation as instruction to keep working. The stop_hook_active guard at the top prevents infinite loops: if Claude is already continuing because of a previous stop-hook block, the script lets it exit. Hooks represent the most automated approach to subagent orchestration. Conversational invocation or CLAUDE.md instructions are the better starting point; hooks come later, as workflows mature. For complete hooks configuration, see Claude Code power user customization: how to configure hooks or our Claude Code hooks docs . Practical patterns for using subagents The following patterns demonstrate subagent direction applied to common scenarios. Research before implementing When adding a feature to unfamiliar code, delegating research to a subagent first keeps the implementation discussion informed rather than exploratory, for

采集自 Claude Blog，由 collect_claude_blog.py 自动采集

Claude Enterprise, now available self-serve¶

来源: Claude Blog 发布日期: Feb 12, 2026 采集时间: 2026-06-01 价值评分: 9.0/10 正文字数: ~2598 字符

摘要¶

Starting today, any organization can purchase Claude Enterprise directly on our website with no sales conversation required. Set up your workspace, configure SSO, and start inviting team members in minutes.

正文内容¶

Starting today, any organization can purchase Claude Enterprise directly on our website with no sales conversation required. Set up your workspace, configure SSO, and start inviting team members in minutes. ‍

Claude Enterprise Claude Enterprise gives your entire organization access to Claude, Claude Code , and Cowork — alongside enterprise security and the ability to work with large codebases, extensive document sets, and broad organizational context. Your teams can brainstorm with Claude, delegate coding tasks from the web and terminal, or hand off complex multi-step projects and get back finished deliverables. With Cowork using plugins , you can turn Claude into a specialist for sales, finance, legal, marketing, and more. Claude also works with the tools your teams already use. Connect to Microsoft 365, Slack, and other platforms through connectors , or collaborate with Claude directly in Excel and PowerPoint through built-in chat sidebars. Claude Enterprise includes the security and administrative controls organizations need to deploy AI at scale: Manage user access with single sign-on (SSO) and domain capture, and automate provisioning through SCIM. Audit logs and a Compliance API give your security and compliance teams visibility into how Claude is being used, and administrators can manage permissions across the organization. Administrators can set custom data retention policies and track team activity, feature adoption, and spend through usage analytics Anthropic does not train models on Claude Enterprise content by default. How organizations use Claude Enterprise With Claude, each team can accelerate their work and transform what individuals and teams can achieve. Sales teams prepare for client meetings by having Claude search through correspondence, calendar context, and the latest prospect information — turning hours of prep into minutes. Engineering teams connect their codebases and use Claude Code to accelerate development across the full software lifecycle. Marketing teams draft, translate, and refine content while maintaining brand consistency. Product managers build interactive prototypes and prioritize roadmaps. Finance teams do deep data analysis, forecasting, and modeling directly in Excel.

Explore more product news and best practices for teams building with Claude.

Transform how your organization operates with Claude

Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.

Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.

采集自 Claude Blog，由 collect_claude_blog.py 自动采集

Introducing Agent Skills¶

来源: Claude Blog 发布日期: Oct 16, 2025 采集时间: 2026-06-01 价值评分: 9.0/10 正文字数: ~1876 字符

摘要¶

Claude can now use Skills to improve how it performs specific tasks. Skills are folders that include instructions, scripts, and resources that Claude can load when needed. Claude will only access a skill when it's relevant to the task at hand.

正文内容¶

Update: We've added organization-wide management for skills , a directory featuring partner-built skills, and published Agent Skills as an open standard for cross-platform portability. (December 18, 2025) Claude can now use Skills to improve how it performs specific tasks. Skills are folders that include instructions, scripts, and resources that Claude can load when needed. Claude will only access a skill when it's relevant to the task at hand. When used, skills make Claude better at specialized tasks like working with Excel or following your organization's brand guidelines.

Claude Code Skills extend Claude Code with your team's expertise and workflows. Install skills via plugins from the anthropics/skills marketplace. Claude loads them automatically when relevant. Share skills through version control with your team. You can also manually install skills by adding them to ~/.claude/skills . The Claude Agent SDK provides the same Agent Skills support for building custom agents. Getting started Claude apps: User Guide & Help Center API developers: Documentation Claude Code: Documentation Example Skills to customize: GitHub repository What's next We're working toward simplified skill creation workflows and enterprise-wide deployment capabilities, making it easier for organizations to distribute skills across teams. Keep in mind, this feature gives Claude access to execute code. While powerful, it means being mindful about which skills you use—stick to trusted sources to keep your data safe. Learn more .

Explore more product news and best practices for teams building with Claude.

Transform how your organization operates with Claude

Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.

Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.

采集自 Claude Blog，由 collect_claude_blog.py 自动采集

Structured outputs on the Claude Developer Platform¶

来源: Claude Blog 发布日期: Nov 14, 2025 采集时间: 2026-06-01 价值评分: 9.0/10 正文字数: ~3135 字符

摘要¶

Structured outputs on the Claude Developer Platform guarantee API responses match your JSON schemas and tool definitions. Now in public beta for Sonnet 4.5 and Opus 4.1, this feature eliminates parsing errors and failed tool calls for production applications.

正文内容¶

Guarantee responses match your JSON schemas and tool definitions with structured outputs.

Update: Now generally available (GA) natively on the Claude Developer Platform and in Amazon Bedrock for Claude Sonnet 4.5, Opus 4.5, and Haiku 4.5. GA adds support for more complex schemas. (Feb 4, 2026) Update: Now available on Claude Haiku 4.5—supported on the Claude Developer Platform, natively and in Microsoft Foundry. (Dec 4, 2025) The Claude Developer Platform now supports structured outputs for Claude Sonnet 4.5 and Opus 4.1. Available in public beta, this feature ensures API responses always match your specified JSON schemas or tool definitions. With structured outputs, developers can eliminate schema-related parsing errors and failed tool calls by ensuring that Claude's responses conform to a defined schema—whether you're extracting data from images, orchestrating agents, or integrating with external APIs. Building reliable applications For developers building applications and agents in production, a single error in data formatting can cause cascading failures. Structured outputs solves this by guaranteeing your response matches the exact structure you define, without any impact to model performance. This makes Claude dependable for applications and agents where accuracy is critical, including: Data extraction when downstream systems rely on error-free, consistent formats. Multi-agent architectures where consistent communication between agents is critical for a performant, stable experience. Complex search tools where multiple search fields must be filled in accurately and conform to specific patterns. Structured outputs can be used two ways: with JSON or tools. When used with JSON, you provide your schema definition in the API request. For tools, you define your tool specifications, and Claude's output conforms to those tool definitions automatically. The end result is a reliable output, reduced retries, and a simplified codebase that no longer needs failover logic or complex error handling. Customer Spotlight: OpenRouter OpenRouter provides 4M+ developers access to all major AI models through a single, unified interface. "Structured outputs have become a really valuable part of the agentic AI stack. Agents constantly ingest and produce structured data, so Anthropic’s structured outputs close a real gap for developers. Agent workflows run reliably, every time, and teams can focus on their customers rather than debugging tool calls,” said Chris Clark, COO, OpenRouter. Getting started Structured outputs is now available in public beta for Sonnet 4.5 and Opus 4.1 on the Claude Developer Platform, with support for Haiku 4.5 coming soon. Explore our documentation for supported JSON schema types, implementation examples, and best practices.

Explore more product news and best practices for teams building with Claude.

Transform how your organization operates with Claude

Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.

Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.

采集自 Claude Blog，由 collect_claude_blog.py 自动采集

Introducing the Claude Team plan and iOS app¶

来源: Claude Blog 发布日期: May 01, 2024 采集时间: 2026-06-01 价值评分: 9.0/10 正文字数: ~1638 字符

摘要¶

Claude now offers a Team plan with increased usage and a free iOS app. Today, we’re announcing two updates for Claude: a new Team plan and an iOS app. The Team plan enables ambitious teams to create a workspace with increased usage for members and tools for managing users and billing.

正文内容¶

Claude now offers a Team plan with increased usage and a free iOS app.

Today, we’re announcing two updates for Claude: a new Team plan and an iOS app. The Team plan enables ambitious teams to create a workspace with increased usage for members and tools for managing users and billing. It’s the best way for teams across industries to leverage our next-generation Claude 3 model family . This plan is available for $30 per user per month. The Claude iOS app is available to download for free for all Claude users. It offers the same intuitive experience as mobile web, including syncing your chat history and support for taking and uploading photos. Claude is designed to help individuals—and now teams—harness the power of the industry’s most advanced AI models. Whether you need a partner for deep work, a knowledgeable expert, a creative collaborator, or an assistant that’s available instantly, Claude augments every employee's capabilities and enables businesses to achieve new levels of productivity to drive better results. Team plan Claude enables companies to shape their workflows based on their teams' unique needs and goals, rather than being limited by their existing tools. Built with security and data privacy in mind, Claude helps protect sensitive business information.

Explore more product news and best practices for teams building with Claude.

Transform how your organization operates with Claude

Product updates, how-tos, community spotlights, and more. Delivered monthly to your inbox.

Please provide your email address if you'd like to receive our monthly developer newsletter. You can unsubscribe at any time.

采集自 Claude Blog，由 collect_claude_blog.py 自动采集