Claude Blog 采集 (2026-07-24)¶
共采集 5 篇文章
📋 文章索引¶
- Building verification loops in Claude Code with skills - Jul 22, 2026 (评分: 9.5)
- How Outtake built a cyber investigator on Claude - Jul 22, 2026 (评分: 9.0)
- Steering Claude Code: when to use CLAUDE.md, skills, hooks, and subagents - Jun 18, 2026 (评分: 9.0)
- The full Claude Desktop experience on AWS, Google Cloud, and Microsoft Foundry - Jun 22, 2026 (评分: 9.0)
- New in Claude Managed Agents: run agents on a schedule and store environment variables in vaults - Jun 09, 2026 (评分: 9.0)
Building verification loops in Claude Code with skills¶
来源: Claude Blog 发布日期: Jul 22, 2026 采集时间: 2026-07-24 价值评分: 9.5/10 正文字数: ~6823 字符
摘要¶
How Anthropic builds verification loops in Claude Code: turn your manual checks into skills so Claude tests, fixes, and verifies its own work.
正文内容¶
How to turn your manual checks into skills, so Claude closes its own feedback loop.
Most agentic coding sessions follow a loop: you ask for a change, Claude gathers context, takes action, verifies the results, and if needed, loops back to gather additional context. Verification is how agents check their work before responding. Claude already does some of this from observing the deterministic signals in your codebase, including type checkers, linters, tests, and runtime errors. Whatever Claude can't infer becomes the steps you take to manually check a feature. These manual steps, however, can be transformed into verification loops. In Claude Code , a verification loop is an iterative process where Claude checks and attempts to fix the work.
Make it a skill The most common way to encode repetitive steps into a verification loop is to write it as a skill , and the fastest way to create a skill is to install the skill-creator plugin and let Claude interview you: Example: /skill-creator Create a skill for verifying frontend changes end-to-end. Interview me about my workflow. You can also hand-write a skill by dropping a markdown file in .claude/skills/ inside your project. The simplest possible verification skill is a few lines of frontmatter plus a body: # .claude/skills/verify-log-hygiene/SKILL.md --- name: verify-log-hygiene description: Check that error logs include the request ID and never include the request body. Use when the diff touches error handling or logging. allowed-tools: [Read, Edit, Grep] --- Read the error-handling paths in the current diff. For each log call on an error path, confirm it includes the request ID and does not pass the request body, headers, or any user-supplied payload. Report each violation with file:line, then fix it: add the request ID where it's missing and strip the payload from the log call. The full schema and the philosophy behind it are in our complete guide to building skills . Match the check to where it runs The next thing to determine will be how the verification loop kicks off: standalone, embedded, chained, or tied to PR. Standalone You invoke it deliberately, after the artifact exists. A standalone skill earns its place for cross-cutting checks that don't apply every time: a pre-commit security scan, a pre-PR accessibility audit, license-header verification across a repo. Anything you want available across many workflows but don't want firing on every code change. The cost is that each invocation is still a turn you have to remember to take. The signal that you've outgrown standalone is when you're running it after every change. At that point, the procedure has earned a permanent home: embed it or chain it. Embedded Fires automatically as part of the producing skill. The check belongs to one specific workflow, and the workflow now runs it without you asking. The simplest version is a one-line append to the producing skill's body: # .claude/skills/scaffold-component/SKILL.md --- name: scaffold-component description: Scaffold a new React component under src/components/, including the component file, its co-located test, and an index export. Use when the user asks to create a new component. allowed-tools: [Read, Write, Edit, Bash, Glob] --- # Scaffold a new React component Given a component name (PascalCase), create the following under src/components/<Name>/: 1. <Name>.tsx: function component with a typed props interface and a default export. 2. <Name>.test.tsx: React Testing Library test that renders the component and asserts it mounts without throwing. 3. index.ts: re-export the default and any named exports. Follow the patterns in src/components/Button/ as the reference. Match the import alias style (@/components/...) used throughout the codebase. # code continues... After creating the component file, run eslint on it and address any errors before reporting completion. Verify the embed works by invoking the skill on a fresh task and confirming the new step runs as part of the output. If it doesn't, the skill's description or earlier instructions aren't pulling the appended check in. Embedded only works on skills you can edit: ones you wrote yourself, or ones installed at a project level where the SKILL.md file is under your control. Built-in skills and plugin-managed skills (the kind that get overwritten on update) are off-limits for this pattern; for those, chain instead. Skip embedded for checks that span workflows; those want standalone, so you can invoke them from any context. Chained One skill calls another at its end, and several verified handoffs run end-to-end. Members of Anthropic's Claude Code team use this pattern in their day-to-day: /code-review hunts for bugs, /simplify cleans up the diff, a /verify skill confirms end-to-end behavior, and a custom /design skill checks against guidelines in a DESIGN.md file if the change touched UI. Chaining is also how you add verification to a skill you can't modify: build a custom wrapper skill that invokes the original, then invokes your verification skill, as depicted below: # .claude/skills/safe-refactor/SKILL.md Run /simplify on the current diff first. When /simplify finishes, invoke /verify-no-public-api-changes. What started as a habit ("I always run /verify after /simplify") becomes a contract ("/simplify always runs /verify when it finishes"). The chain runs the whole dev cycle on its own. You only step in when something escalates back to you. You can skip chaining when the steps are independent enough that you sometimes want to run one without the others; chaining trades flexibility for automation. Chained verification loops can increase token spend, so it's best to test these loops before deploying them broadly. On every PR Once the chain is solid for your own changes, the same procedure can run on every PR. A teammate's change passes the same gates yours did, whether they remembered to invoke the chain or not. The infrastructure is the same kind of thing as the chain you already wrote, one step further along: the same skills, the same rubrics, the same standards, applied without depending on the author's diligence. This is where verification stops being personal infrastructure and becomes team infrastructure . The check you wrote down to save yourself two minutes a week is now saving everyone two minutes a week, on every change. Hold off on PR-wide gates while the chain is still in flux; every adjustment becomes a team-visible event.
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
How Outtake built a cyber investigator on Claude¶
来源: Claude Blog 发布日期: Jul 22, 2026 采集时间: 2026-07-24 价值评分: 9.0/10 正文字数: ~1252 字符
摘要¶
How Outtake ensures multi-hour agent sessions stay on track to uncover attack network operations
正文内容¶
How Outtake ensures multi-hour agent sessions stay on track to uncover attack network operations
In our series, How startups build with Claude, we highlight how startups are transforming their industries with AI. In this article, we share how Outtake built an autonomous cyber investigator that detects, investigates, and dismantles digital threats, from cloned login pages to entire adversarial networks. The quick pitch Name Outtake Founded 2023 Founders Alex Dhillon (CEO), formerly of Palantir's moonshot team Growth Grew annual recurring revenue 6x and its customer base more than 10x year-over-year, scanning 20M+ potential cyberattacks in 2025 alone.
How Outtake built a complex long-running agent with Claude Outtake built the Recon Agent in roughly four stages. Each stage was about understanding what a good investigation looked like, then progressively handing that judgment to the agent.
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
Steering Claude Code: when to use CLAUDE.md, skills, hooks, and subagents¶
来源: Claude Blog 发布日期: Jun 18, 2026 采集时间: 2026-07-24 价值评分: 9.0/10 正文字数: ~1792 字符
摘要¶
Seven ways to steer Claude Code—CLAUDE.md files, rules, skills, subagents, hooks, and more—and when to use each, based on context cost and authority.
正文内容¶
Rules Rules are markdown files in .claude/rules/ that give Claude specific constraints or conventions. Unscoped rules behave like CLAUDE.md in that they are always loaded at session start and get re-injected on compaction. This can waste tokens by loading context even when it's not relevant for the task at hand. Path-scoped rules allow you to load rule instructions only when they are relevant by adding a paths field that controls when they load. For example: a rule scoped to src/api/ stays out of context during a docs-only session. It would only be loaded whenever Claude reads files within that src/api/ directory. Here’s what that looks like: --- paths: - "src/api/" - "*/.handler.ts" --- All API handlers must validate input with Zod before processing. Tip : A file-specific constraint, like "migrations are append-only," fits best as a rule placed in your paths: frontmatter. Reach for a path scoped rule over a nested CLAUDE.md file when the instruction regards a cross-cutting concern or file that appears in multiple (but not all) corners of the codebase. Skills Skills live in .claude/skills/ as folders of instructions, scripts, and resources that Claude loads dynamically. Each skill has a SKILL.md file with a name, description, and body. Only the name and description load at session start; the full body loads when Claude invokes the skill, either through a slash command (/code-review) or by auto-matching the task.
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
The full Claude Desktop experience on AWS, Google Cloud, and Microsoft Foundry¶
来源: Claude Blog 发布日期: Jun 22, 2026 采集时间: 2026-07-24 价值评分: 9.0/10 正文字数: ~3517 字符
摘要¶
Deploy the full Claude desktop experience - chat, Claude Cowork, and Claude Code - using inference on AWS, Google Cloud and Microsoft Foundry. Available today in beta.
正文内容¶
Organizations that use Claude Desktop through AWS, Google Cloud, and Microsoft Foundry now get the full Desktop experience — chat, Claude Cowork, and Claude Code, all in one app. Now IT teams can keep inference inside their own environment across products, and deploy Claude Desktop organization-wide with per-user SSO, MDM policy templates, an offline installer option, and an M365 connector that can run entirely on the device. Inference runs on your cloud in the regions you configure and conversation history is stored locally. You control the endpoints data connectors reach and the aggregated telemetry Anthropic receives. One surface for the entire organization Until today, customers using Claude Desktop through AWS, Google Cloud, and Microsoft Foundry only had access to Claude Cowork and Claude Code. Now, one deployment covers every role, and each surface has its own policy key, so you decide who gets what, and when. Chat for quick answers and thinking through a problem. Claude Cowork for the work your people would rather hand off: Claude researches across approved sources, works with the files already on the device and builds the deliverable, surfacing results when it’s done. Claude Code for engineers who want agentic coding without living in a terminal. Deployment controls Deploying Claude Desktop organization-wide means working within the systems you already have. Sign in like any work app. Employees use the same work account they use for everything else: IAM Identity Center, Workforce Identity Federation, Microsoft Entra ID, or any OIDC provider like Okta. No shared keys to rotate, no cloud credentials on end-user machines. Deploy like any app you already manage. Export policy templates from the setup UI and push them through Intune, GPO, or Jamf. An offline installer covers air-gapped environments. Know it works before anyone sees it. Test every connector, confirm which Claude models your provider serves, and verify the connection, all before rollout. A model guard keeps routing on Claude, including in GovCloud, even if a setting is misconfigured. Start small, expand as adoption grows. Chat, Claude Cowork, and Claude Code each have their own policy key, so you can give non-technical teams chat and Claude Cowork, engineering Claude Code, and then broaden access as teams adopt each surface. Your hard-deny rules apply across every tab. Bring Claude to where the work lives. A Microsoft 365 connector gives Claude access to mail and documents through your own Entra app, with tenant allowlisting and beta support for GCC High/DoD endpoints. For the strictest residency requirements, use our local connector, and the connection stays between the device and Microsoft. "We rolled out Claude Desktop fast through our existing cloud environment — no separate vendor contract. Our own LLM Gateway let one team deploy it to hundreds of users worldwide, with no heavy infrastructure build-out." - Sarang Oh, Analytics/AI Team Leader, Hanwha Solutions Getting started For admins, the deployment guide walks through SSO, policy templates, and pre-rollout validation. Or contact your account team and we'll help you plan the rollout.
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
New in Claude Managed Agents: run agents on a schedule and store environment variables in vaults¶
来源: Claude Blog 发布日期: Jun 09, 2026 采集时间: 2026-07-24 价值评分: 9.0/10 正文字数: ~1191 字符
摘要¶
Claude Managed Agents now supports scheduled deployments and vaults: run agents on a cron schedule and securely authenticate CLI tools and other services.
正文内容¶
Starting today, Claude Managed Agents can run on a schedule and securely access CLI tools and other authenticated services. Both features are now available in public beta on the Claude Platform. Scheduled deployments: Run agents on a schedule Agents can now run on a schedule, completing routine work automatically. A scheduled deployment runs a Claude Managed Agent on a cron schedule. Each time the schedule fires, the agent starts a new session and completes its task, with no scheduler for you to build or host. Use it for recurring work like a nightly data sync, a weekly compliance scan, or a daily digest. Once a deployment is live, you can pause, resume, or archive it at any time, or trigger additional runs on demand.
Getting started Explore our documentation to learn more or visit the Claude Console to deploy your first agent.
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