Use HTTP correctly with proper methods, status codes, headers, and caching.
数据来源:ClawHub。 在 ClawSkills 查看
选择你使用的 Agent
方法一:命令行安装(推荐)
推荐(无需提前安装 clawhub)
npx clawhub@latest --dir ~/.claude/skills install http或使用 clawhub CLI(需提前安装)
clawhub --dir ~/.claude/skills install http⚠️ 需要 Node.js 18+,没有 Node?请使用下方方法二直接下载 ZIP。 安装 Node.js →
方法二:手动下载安装(无需 Node)
下载 ZIP,解压后将文件夹放到以下路径,重启 Agent 即可:
安装路径
~/.claude/skills/http/💡解压后将文件夹放到上方路径,重启 Agent 即可生效
--- name: HTTP description: Use HTTP correctly with proper methods, status codes, headers, and caching. metadata: {"clawdbot":{"emoji":"🌐","os":["linux","darwin","win32"]}} ---
Location header with absolute URL—relative may fail in older clientsCache-Control: no-store for sensitive data—never written to diskno-cache still caches but revalidates every time—not "don't cache"private, max-age=0, must-revalidate for user-specific, always-fresh contentpublic, max-age=31536000, immutable for versioned static assetsVary: Accept-Encoding, Authorization when response depends on these headers—forgetting Vary breaks cachingETag + If-None-Match: prefer for APIs—content hash based"abc" vs W/"abc"—weak allows semantically equivalent responsesIf-Match for optimistic locking: fail update if resource changed since readIf-Match fails—not 409 ConflictAccess-Control-Max-Age—set to 86400 to reduce OPTIONS spamStrict-Transport-Security: max-age=31536000; includeSubDomains—HSTS, once set can't easily undoX-Content-Type-Options: nosniff—prevents MIME sniffing attacksX-Frame-Options: DENY or SAMEORIGIN—prevents clickjackingContent-Security-Policy—complex but essential; start with report-only modeAccept-Ranges: bytes signals support—clients can request partial contentRange: bytes=0-1023 requests first 1024 bytes; bytes=-500 requests last 500Content-Range: bytes 0-1023/5000Content-Range: bytes */5000{"error": {"code": "VALIDATION_FAILED", "message": "...", "details": [...]}}Idempotency-Key: Retry-After header—can be seconds or HTTP dateVary: must include headers that affect response—CORS without Vary: Origin breaksContent-Disposition: attachment; filename="report.pdf" for downloadsX-Request-ID: generate if not present, propagate to downstream servicesAccept-Language for localized responses—respect with graceful fallbackContent-Length or chunked = connection close after responseTransfer-Encoding: chunked for streaming—can't set Content-LengthConnection: Upgrade, Upgrade: websocket安装 HTTP 后,可以对 AI 说这些话来触发它
Help me get started with HTTP
Explains what HTTP does, walks through the setup, and runs a quick demo based on your current project
Use HTTP to use HTTP correctly with proper methods, status codes, headers, and ...
Invokes HTTP with the right parameters and returns the result directly in the conversation
What can I do with HTTP in my marketing & growth workflow?
Lists the top use cases for HTTP, with example commands for each scenario
将技能文件夹放到 ~/.claude/skills/http/ 目录(个人级,所有项目可用),或 .claude/skills/http/(项目级)。重启 AI 客户端后,用 /http 主动调用,或让 AI 根据上下文自动发现并使用。
HTTP 支持 Claude、Cursor、OpenClaw,可与这些 AI 平台无缝集成,扩展其能力。
HTTP 可免费安装使用。请查阅仓库了解许可证信息。
Use HTTP correctly with proper methods, status codes, headers, and caching.
HTTP 属于「Marketing & Growth」分类,该分类的技能帮助 AI 智能体在此领域执行专业任务。
Automate my marketing & growth tasks using HTTP
Identifies repetitive steps in your workflow and sets up HTTP to handle them automatically