Token导航 LogoToken导航

Agent Skill · 研究检索

box(盒子)

box 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。

box 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 40,677。

总安装 1,261周安装 51GitHub Stars 4下载 396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:box(盒子)
来源仓库:https://github.com/box/box-for-ai
仓库路径:skills/box
安装命令:
npx skills add https://github.com/box/box-for-ai --skill box
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/box/box-for-ai --skill box

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。

Box 用于在应用代码中实现内容工作流,支持搜索、AI 集成和元数据管理。适合需要与 Box 平台集成进行内容检索、协作或批量操作的场景。根据操作类型选择工具,优先使用 MCP 结构化接口处理常见任务。使用前需完成工具清单并确认身份认证方式,避免重复造轮子。

涉及共享、webhook 或 AI 功能时,应先测试最小可行路径再扩展。

来源仓库内路径:skills/box

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/box/box-for-ai --skill box;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

box 怎么安装?

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

box 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。

box 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。

box 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。

box 用的人多吗?

公开数据是累计安装 1,261 次,近一周安装 51 次,下载 396 次,GitHub 4 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。

box 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
box

盒子 · 本页

1,2614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

Box

Overview

Implement Box content workflows in application code. Reuse the repository's existing auth and HTTP or SDK stack whenever possible, identify the acting Box identity before coding, and make the smallest end-to-end path work before layering on sharing, metadata, webhooks, or AI.

Route The Request

Tool selection

After completing step 0 (tool inventory), use this table to pick the right tool for the operation:

Operation typePreferRationale
Most agent workflows (search, AI, content management, metadata, hubs)MCPStructured I/O, concurrent-safe, covers the common cases
Bulk operations (batch moves, folder trees, batch metadata)CLICompact output, --fields filtering, full API surface without requiring manual REST auth
Verification and smoke testsCLIReproducible, user can copy-paste commands
Operations outside MCP scopeCLIFull API coverage
Last-resort fallback when MCP is unavailable and CLI is unavailable or not an optionDirect RESTOnly after explicit user confirmation and REST auth setup guidance
Building application code (SDK/REST endpoints, webhook handlers)SDK or REST in codeNot agent tooling — write code the user ships

MCP covers the majority of common agent workflows and is the default when it has a matching tool. Use CLI when the operation falls outside MCP's scope, when compact and field-filtered output matters, or for reproducible verification commands. If MCP is unavailable, guide the user through MCP setup first; if CLI is unavailable, guide the user through CLI setup next. Use direct REST only as a last resort after explicitly asking the user to confirm that REST fallback is acceptable.

Domain routing

Choose which reference files to read based on what the user needs:

If the user needs...Read firstPair withMinimal verification
Uploads, folders, listings, downloads, shared links, collaborations, or metadatareferences/content-workflows.mdreferences/auth-and-setup.mdRead-after-write call using the same actor
Organizing, reorganizing, or batch-moving files across folders; bulk metadata tagging; migrating folder structuresreferences/bulk-operations.mdreferences/content-workflows.md, references/auth-and-setup.md, references/ai-and-retrieval.mdInventory source, verify move count matches plan
Event-driven ingestion, new-file triggers, or webhook debuggingreferences/webhooks-and-events.mdreferences/auth-and-setup.md, references/troubleshooting.mdSignature check plus duplicate-delivery test
Search, document retrieval, summarization, extraction, or Box AIreferences/ai-and-retrieval.mdreferences/auth-and-setup.mdRetrieval-quality check before answer formatting
401, 403, 404, 409, 429, missing content, or wrong-actor bugsreferences/troubleshooting.mdreferences/auth-and-setup.mdReproduce with the exact actor, object ID, and endpoint
Unsure which workflow appliesreferences/workflows.mdreferences/auth-and-setup.mdChoose the smallest Box object/action pair first

Workflow

Follow these steps in order when coding against Box.

  1. Inventory available Box tooling:

- MCP: Call who_am_i. If it fails, try mcp_auth. If auth still fails, read references/auth-and-setup.md for MCP setup steps. Record whether MCP is available. - CLI: Run box users:get me --json. Record whether CLI is available. - If MCP is unavailable, walk the user through MCP setup and retry MCP auth before considering other tooling. - If CLI is unavailable, walk the user through CLI setup and retry box users:get me --json. - If MCP remains unavailable and CLI remains unavailable or the user declines CLI, ask for explicit confirmation before using direct REST fallback. If approved, use references/rest-calls.md for auth and request patterns. - If the task is building application code (adding SDK endpoints, webhook handlers), tooling availability is secondary — proceed to step 1.

  1. Inspect the repository for existing Box auth, SDK or HTTP client, env vars, webhook handlers, Box ID persistence, and tests.
  2. Determine the acting identity before choosing endpoints: connected user, enterprise service account, app user, or platform-provided token.
  3. Select the tool using the tool selection table and identify the domain reference using the domain routing table above.
  4. Confirm whether the task changes access or data exposure. Shared links, collaborations, auth changes, large-scale downloads, and broad AI retrieval all need explicit user confirmation before widening access or scope.
  5. Read the reference for the selected tool (references/mcp-tool-patterns.md for MCP, references/box-cli.md for CLI, references/rest-calls.md for direct REST fallback) and the domain reference from the routing table:

- Box MCP tool usage patterns: references/mcp-tool-patterns.md - Box CLI local verification: references/box-cli.md - Direct REST fallback patterns: references/rest-calls.md - Auth setup, actor selection, SDK vs REST: references/auth-and-setup.md - Workflow router: references/workflows.md - Content operations: references/content-workflows.md - Bulk file organization, batch moves, folder restructuring: references/bulk-operations.md - Webhooks and events: references/webhooks-and-events.md - AI and retrieval: references/ai-and-retrieval.md - Debugging and failure modes: references/troubleshooting.md

  1. Implement the smallest end-to-end flow that proves the integration works.
  2. Add a runnable verification step. Prefer the repository's tests first; otherwise use native Box CLI commands when CLI is available and authenticated. Use direct Box REST verification only as a last resort after explicit user confirmation.
  3. Summarize the deliverable with auth context, Box IDs, env vars or config, and the exact verification command or test.

Guardrails

  • Preserve the existing Box auth model unless the user explicitly asks to change it.
  • Check the current official Box docs before introducing a new auth path, changing auth scope, or changing Box AI behavior.
  • Prefer an official Box SDK when the codebase already uses one or the target language has a maintained SDK. Otherwise use direct REST calls with explicit request and response handling.
  • In agent workflows, do not jump straight to direct REST when MCP or CLI can be set up. Offer setup guidance for MCP first and CLI second before proposing REST fallback.
  • Never use direct REST fallback silently. Ask the user for explicit confirmation before proceeding with REST calls.
  • Keep access tokens, client secrets, private keys, and webhook secrets in env vars or the project's secret manager.
  • Distinguish file IDs, folder IDs, shared links, metadata template identifiers, and collaboration IDs.
  • Treat shared links, collaborations, and metadata writes as permission-sensitive changes. Confirm audience, scope, and least privilege before coding or applying them.
  • Require explicit confirmation before widening external access, switching the acting identity, or retrieving more document content than the task truly needs.
  • When a task requires understanding document content — classification, extraction, categorization — use Box AI (Q&A, extract) as the first method attempted. Box AI operates server-side and does not require downloading file bodies. Fall back to metadata inspection, previews, or local analysis only if Box AI is unavailable, not authorized, or returns an error on the first attempt.
  • Pace Box AI calls at least 1–2 seconds apart. For content-based classification of many files, classify a small sample first to validate the prompt and discover whether cheaper signals (filename, extension, metadata) can sort the remaining files without additional AI calls.
  • Avoid downloading file bodies or routing content through external AI pipelines when Box-native methods (Box AI, search, metadata, previews) can answer the question server-side.
  • Request only the fields the application actually needs, and persist returned Box IDs instead of reconstructing paths later.
  • Run Box CLI commands strictly one at a time. The CLI does not support concurrent invocations and parallel calls cause auth conflicts and dropped operations. For bulk work in agent-driven sessions, default to CLI and use REST only after MCP/CLI setup attempts fail or CLI is not an option and the user explicitly confirms REST fallback.
  • Make webhook and event consumers idempotent. Box delivery and retry paths can produce duplicates.
  • Keep AI retrieval narrow for search and Q&A tasks. Search and filter first, then retrieve only the files needed for the answer. This does not apply to Box AI classification — when classifying documents, Box AI should be tried first per the content-understanding guardrail above.
  • Do not use box configure:environments:get --current as a routine auth check because it can print sensitive environment details.

Verification

  • Prefer the repository's existing tests or app flows when they already cover the changed Box behavior.
  • If no better verification path exists, prefer native box CLI commands when box is installed and authenticated.
  • Use direct REST verification only after confirming MCP and CLI are unavailable or not an option and after the user explicitly approves REST fallback.
  • For REST fallback, guide the user through token setup (BOX_ACCESS_TOKEN) and safe auth handling before issuing requests.
  • Confirm CLI auth with box users:get me --json.
  • Verify mutations with a read-after-write call using the same actor, and record the object ID.
  • For webhooks, test the minimal happy path, duplicate delivery, and signature failure handling.
  • For AI flows, test retrieval quality separately from answer formatting.

Example smoke checks:

box users:get me --json
box folders:get 0 --json --fields id,name,item_collection
box folders:items 0 --json --max-items 20
box search "invoice" --json --limit 10
curl -sS -H "Authorization: Bearer $BOX_ACCESS_TOKEN" -H "Accept: application/json" "https://api.box.com/2.0/folders/0?fields=id,name,item_collection"

Deliverable

The final answer should include:

  • Acting auth context used for the change
  • Box object type and IDs touched
  • Env vars, secrets, or config expected by the integration
  • Files or endpoints added or changed
  • Exact verification command, script, or test path
  • Any permission-sensitive assumptions that still need confirmation

References

  • references/mcp-tool-patterns.md: best-practice patterns for working with Box content via the Box MCP server — search, file writes, metadata extraction, Box AI tool selection, and general guidelines
  • references/auth-and-setup.md: auth path selection, SDK vs REST choice, existing-codebase inspection, and current Box doc anchors
  • references/box-cli.md: CLI-first local auth, smoke-test commands, and safe verification patterns
  • references/rest-calls.md: direct REST fallback patterns, auth setup, and safe request templates
  • references/workflows.md: quick workflow router when the task is ambiguous
  • references/content-workflows.md: uploads, folders, listings, downloads, shared links, collaborations, metadata, and file moves
  • references/bulk-operations.md: organizing files at scale, batch moves, folder hierarchy creation, serial execution, and rate-limit handling
  • references/webhooks-and-events.md: webhook setup, event-feed usage, idempotency, and verification
  • references/ai-and-retrieval.md: search-first retrieval, Box AI usage, and external AI guardrails
  • references/troubleshooting.md: common failure modes and a debugging checklist
  • examples/box-prompts.md: example prompts for realistic use cases

平台分布

Codex

37.64%
按下载量换算149

Claude

30.91%
按下载量换算122

Cursor

17.15%
按下载量换算68

Gemini CLI

9.19%
按下载量换算36

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