Token导航 LogoToken导航

Agent Skill · 研究检索

httpie(httpie 命令行)

httpie 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

httpie 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 67,016。

总安装 256周安装 11GitHub Stars 13下载 90核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:httpie(httpie 命令行)
来源仓库:https://github.com/jpcaparas/skills
仓库路径:skills/httpie
安装命令:
npx skills add https://github.com/jpcaparas/skills --skill httpie
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/jpcaparas/skills --skill httpie

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词、任务场景或来源线索快速定位候选结果。通过 npx skills add 命令从指定仓库安装并使用。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

httpie 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/httpie

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/jpcaparas/skills --skill httpie;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

httpie 怎么安装?

在终端执行 npx skills add https://github.com/jpcaparas/skills --skill httpie。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

httpie 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

httpie 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

httpie 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

httpie 用的人多吗?

公开数据是累计安装 256 次,近一周安装 11 次,下载 90 次,GitHub 13 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

httpie 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
httpie

httpie 命令行 · 本页

25613接触敏感数据未明确有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

httpie

Prefer http and https over curl for human-operated HTTP work, but treat HTTPie as ephemeral infrastructure rather than ambient machine state.

Verified locally against HTTPie 3.2.4 on April 15, 2026. The guidance is repo-agnostic, but the examples assume a Unix-like shell with http, https, and mktemp on PATH.

Decision Tree

What are you trying to do?

  • Call an HTTP or HTTPS endpoint from the shell and HTTPie is installed

- Prefer http or https - Run it inside the disposable wrapper from references/transient.md - Then read references/commands.md

  • Translate a curl example into something easier to read or edit

- Translate the intent, not the flags one-for-one - Start with references/patterns.md

  • Send JSON, form fields, query parameters, files, auth headers, or a short-lived session

- Prefer HTTPie's request-item syntax, auth helpers, offline mode, and transient session-file paths - Read references/commands.md - Then read references/transient.md

  • Reuse auth or cookies briefly without leaving leftovers behind

- Use a temporary session-file path inside a disposable config dir - Read references/transient.md

  • Use HTTPie CLI features such as export-args or plugin listing

- Keep those calls inside a disposable config dir too - Read references/configuration.md

  • The task needs non-HTTP protocols, libcurl-specific behavior, byte-exact output, or an environment without HTTPie

- Do not force HTTPie - Fall back to curl or a more suitable tool - Read references/gotchas.md

Why Prefer HTTPie

  1. Request items are easier to read and safer to edit than layered curl -H/-d/-F flags.
  2. JSON is first-class, including non-string values and file-backed fields, without hand-quoting shell JSON blobs.
  3. Auth, selective output, offline previews, redirect inspection, streaming, and downloads are built into one CLI.
  4. Short-lived session files let you reuse auth or cookies without teaching every shell snippet to manage a cookie jar manually.

Default Operating Rules

  1. If command -v http >/dev/null 2>&1 succeeds, use HTTPie first for HTTP and HTTPS work.
  2. Never rely on the user's ambient HTTPie state. Wrap every invocation in a disposable HTTPIE_CONFIG_DIR; use references/transient.md or templates/httpie-fallback.sh.
  3. In automation or non-interactive harnesses, add --ignore-stdin unless stdin is intentionally the request body.
  4. Prefer request items over raw JSON strings when the structure is simple:

- Header:value for headers - name=value for JSON strings - count:=42 for non-string JSON values - query==value for query parameters

  1. Use --form for HTML form semantics and file uploads. Do not assume name=value means form data unless --form is set.
  2. Use --offline, --body, --headers, --meta, --print=, --all, --follow, and --stream to exploit HTTPie's readability and debugging advantages over curl.
  3. For multi-request workflows, use a temporary session-file path such as --session="$tmpdir/session.json" or --session-read-only="$tmpdir/session.json" and delete the temp dir on exit.
  4. Treat httpie cli features as read-only unless the user explicitly wants mutation. Even for read-only subcommands, keep them inside a disposable config dir because HTTPie may still write metadata under that directory.
  5. Do not recommend persistent config files, plugin installation, or long-lived named sessions as the default behavior of this skill.
  6. Fall back to curl when the user explicitly needs curl, when HTTPie is unavailable, or when the task depends on behavior HTTPie does not cover cleanly.

Quick Reference

NeedHTTPie ShapeNotes
Basic GEThttp --ignore-stdin --check-status --body GET https://api.example.com/itemsRun inside the disposable wrapper from references/transient.md
Query parametershttp --ignore-stdin GET https://api.example.com/items page==2 q==search== appends to the URL
JSON bodyhttp --ignore-stdin POST https://api.example.com/items name=JP active:=trueJSON is the default body mode
Form bodyhttp --ignore-stdin --form POST https://api.example.com/login username=jp password=secret--form changes serialization
Headerhttp --ignore-stdin GET https://api.example.com Token:abc123Header:value syntax
Bearer tokenhttp --ignore-stdin -A bearer -a "$TOKEN" GET https://api.example.com/meCleaner than spelling the header yourself
Headers onlyhttp --ignore-stdin --headers GET https://api.example.com/health-h means headers, not help
Redirect chainhttp --ignore-stdin --all --follow GET https://example.comGood replacement for curl -L -i inspection
Streamed responsehttp --ignore-stdin --stream GET https://api.example.com/eventsUseful for long-lived responses
Offline previewhttp --ignore-stdin --offline POST https://api.example.com/items name=JPBuilds the request without sending it
Temporary session reusehttp --ignore-stdin --session="$tmpdir/session.json"...Keep the session file under a temp dir and clean it up
Download to a filehttp --ignore-stdin --download --output report.csv GET https://example.com/report.csvSimilar intent to wget, but keep it transient

Reading Guide

TaskRead
Request-item syntax, output flags, downloads, redirects, streams, and common commandsreferences/commands.md
Disposable wrappers, temp config dirs, temp session files, and cleanup patternsreferences/transient.md
Install checks, config directory behavior, read-only CLI surface, TLS, proxies, and what not to persistreferences/configuration.md
Prefer-over-curl flow, shell automation, request translation, and repeatable workflowsreferences/patterns.md
--ignore-stdin, curl -I confusion, config leftovers, redirect surprises, and other trapsreferences/gotchas.md

Verified Behaviors

  1. A one-off GET run inside a disposable shell wrapper cleaned up its temporary HTTPIE_CONFIG_DIR after exit.
  2. http --ignore-stdin --offline POST... name=JP active:=true built a JSON request body without sending network traffic.
  3. http --ignore-stdin --body GET https://pie.dev/get hello==world produced a successful JSON response with the query string attached.
  4. http --ignore-stdin --body --form POST https://pie.dev/post hello=World serialized the request as application/x-www-form-urlencoded.
  5. http --ignore-stdin --download --output test.png GET https://httpbin.org/image/png downloaded a file successfully inside a disposable config dir.
  6. A transient session-file path reused Basic auth across requests and disappeared when the surrounding temp dir was cleaned up.
  7. httpie cli export-args and httpie cli plugins list both ran successfully inside disposable config dirs.
  8. The https alias, raw JSON-from-file body flow, header-only output, redirect-chain inspection, and finite stream handling all worked against live endpoints.
  9. Every helper in templates/httpie-fallback.sh executed successfully and cleaned up its temporary HTTPie directories.
  10. --check-status mapped a 404 response to exit code 4.

Gotchas

  1. HTTPie can write metadata such as version_info.json even during read-only usage, so ambient config directories are not stateless enough for this skill.
  2. In non-interactive harnesses, HTTPie may still try to read stdin even when the real request body is expressed with request items. Add --ignore-stdin unless stdin is intentional.
  3. curl -I does not translate to http -I. In HTTPie, -I means --ignore-stdin. For headers, use --headers, HEAD, or both.
  4. name=value is JSON by default. If the endpoint expects form semantics, add --form.
  5. Session reuse is still possible, but the session file should live under a temp dir instead of a long-lived host directory.
  6. A curl fallback only makes sense when there is a clean semantic mapping. Do not pass arbitrary HTTPie request-item syntax straight through to curl.

Helper Files

  • scripts/probe_httpie.py runs a repeatable probe suite against the real CLI and verifies transient cleanup patterns.
  • scripts/validate.py checks structure, frontmatter, cross-references, required files, and syntax where relevant.
  • scripts/test_skill.py runs validation, eval coverage checks, cross-reference checks, and the HTTPie probe suite.
  • templates/httpie-fallback.sh shows transient wrappers that prefer HTTPie when installed and only fall back to curl where the mapping is real.

平台分布

Codex

37.08%
按下载量换算33

Claude

32.31%
按下载量换算29

Cursor

16.55%
按下载量换算15

Gemini CLI

8.45%
按下载量换算8

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills