Token导航 LogoToken导航

Agent Skill · 研究检索

shelv(shelv 搜索)

shelv 用于查找、检索和筛选相关信息,适合在 OpenClaw 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。

shelv 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 7,801。

总安装 18,010周安装 758GitHub Stars 2下载 6,307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
MIT-0
版本
1.0.3

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:shelv(shelv 搜索)
来源仓库:https://github.com/joshuaohanlon/shelv
安装命令:
openclaw skills install shelv
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install shelv

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。

将 PDF 转换为结构化 Markdown 文件系统,并将它们合并到您的工作空间中,以便使用标准 Unix 工具进行探索

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 OpenClaw;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 openclaw skills install shelv;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

shelv 怎么安装?

在终端执行 openclaw skills install shelv。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

shelv 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。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。

shelv 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。

shelv 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

shelv 用的人多吗?

公开数据是累计安装 18,010 次,近一周安装 758 次,下载 6,307 次,GitHub 2 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。

shelv 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

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。
  • 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 OpenClaw,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。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

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
shelv

shelv 搜索 · 本页

18,0102接触敏感数据MIT-0通过

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

name
shelv
description
Convert PDFs into structured Markdown filesystems and hydrate them into your workspace for exploration with standard Unix tools
version
1.0.3
metadata
openclaw
requires
env
[SHELV_API_KEY]
bins
[curl, tar, jq, shasum]
primaryEnv
SHELV_API_KEY
emoji
📚
homepage
https://shelv.dev
os
[macos, linux]

Shelv

Shelv converts PDF documents (contracts, books, research papers, regulations) into structured Markdown filesystems. Upload a PDF, wait for processing, then hydrate the result into your workspace as real files you can explore with ls, cat, grep, and find.

API base URL: https://api.shelv.dev Auth: Authorization: Bearer $SHELV_API_KEY on every request.

Get your API key at shelv.dev → Settings → API Keys.

Core Workflows

1. Upload a document

Upload a PDF to create a new shelf. Processing runs asynchronously.

SHELF_ID=$({baseDir}/scripts/shelv-upload.sh /path/to/document.pdf --name "My Document")

With options:

# Use a structuring template
SHELF_ID=$({baseDir}/scripts/shelv-upload.sh document.pdf --name "Q4 Contract" --template legal-contract)

# Enable review mode (pause before finalizing)
SHELF_ID=$({baseDir}/scripts/shelv-upload.sh document.pdf --review)

# Upload and wait for processing to complete
SHELF_ID=$({baseDir}/scripts/shelv-upload.sh document.pdf --wait)

The script prints the shelf public ID (e.g. shf_0123456789abcdef01234567) to stdout.

Available templates: book, legal-contract, academic-paper. Omit to let Shelv auto-detect structure.

Inline alternative (without the script):

curl -X POST "https://api.shelv.dev/v1/shelves" \
  -H "Authorization: Bearer $SHELV_API_KEY" \
  -F "file=@document.pdf" \
  -F "name=My Document"

Response (201):

{
  "publicId": "shf_0123456789abcdef01234567",
  "name": "My Document",
  "status": "uploading",
  "template": null,
  "reviewMode": false,
  "pageCount": null,
  "createdAt": "2025-01-15T10:30:00.000Z",
  "updatedAt": "2025-01-15T10:30:00.000Z"
}

2. Poll for processing completion

Wait for a shelf to finish processing:

{baseDir}/scripts/shelv-poll-status.sh shf_0123456789abcdef01234567

The script polls GET /v1/shelves/{id} every 5 seconds and prints the current status. It exits 0 when the shelf reaches ready or review, and exits 1 on failed (with error details) or timeout (10 minutes).

Processing flow: uploading → parsing → structuring → verifying → ready

If review mode is enabled, the flow pauses at review instead of ready.

Inline alternative:

curl -s "https://api.shelv.dev/v1/shelves/$SHELF_ID" \
  -H "Authorization: Bearer $SHELV_API_KEY" | jq '.status'

3. Hydrate a shelf into the workspace

Download and extract the shelf's Markdown filesystem into your workspace:

{baseDir}/scripts/shelv-hydrate.sh shf_0123456789abcdef01234567

This downloads the archive, verifies its checksum, and extracts it to ~/.openclaw/workspace/shelves/<name>/. The script prints a file listing after extraction.

Override the directory name:

{baseDir}/scripts/shelv-hydrate.sh shf_0123456789abcdef01234567 --name my-contract

Replace an existing shelf (required if the directory already exists):

{baseDir}/scripts/shelv-hydrate.sh shf_0123456789abcdef01234567 --force

After hydration, explore the files:

ls ~/.openclaw/workspace/shelves/my-contract/
cat ~/.openclaw/workspace/shelves/my-contract/README.md
find ~/.openclaw/workspace/shelves/my-contract/ -name "*.md"
grep -r "force majeure" ~/.openclaw/workspace/shelves/my-contract/

4. List and browse existing shelves

List all shelves:

curl -s "https://api.shelv.dev/v1/shelves?page=1&limit=20" \
  -H "Authorization: Bearer $SHELV_API_KEY" | jq '.data[] | {publicId, name, status}'

Get the file tree (flat JSON map of path → content):

curl -s "https://api.shelv.dev/v1/shelves/$SHELF_ID/tree" \
  -H "Authorization: Bearer $SHELV_API_KEY" | jq '.files | keys[]'

Response shape:

{
  "shelfPublicId": "shf_0123456789abcdef01234567",
  "name": "My Contract",
  "fileCount": 8,
  "files": {
    "README.md": "# My Contract\
...",
    "clauses/force-majeure.md": "# Force Majeure\
..."
  }
}

5. Read individual files without hydration

Read a single file by path (useful for targeted lookups without downloading the full archive):

curl -s "https://api.shelv.dev/v1/shelves/$SHELF_ID/files/README.md" \
  -H "Authorization: Bearer $SHELV_API_KEY"
curl -s "https://api.shelv.dev/v1/shelves/$SHELF_ID/files/clauses/force-majeure.md" \
  -H "Authorization: Bearer $SHELV_API_KEY"

Returns raw Markdown content (text/markdown).

Workspace Conventions

Hydrated shelves live at:

~/.openclaw/workspace/shelves/{name}/

The {name} is derived from the shelf's display name (lowercased, spaces and special characters replaced with hyphens). Override with --name when hydrating.

If a directory already exists at the target path, the script will refuse to overwrite it unless --force is passed.

After hydration, use standard Unix tools to explore:

# List all files
find ~/.openclaw/workspace/shelves/{name}/ -type f

# Read a specific file
cat ~/.openclaw/workspace/shelves/{name}/README.md

# Search across all files
grep -r "keyword" ~/.openclaw/workspace/shelves/{name}/

# Count files
find ~/.openclaw/workspace/shelves/{name}/ -type f | wc -l

Async Operations

Shelf processing is asynchronous. After uploading, the shelf progresses through:

uploading → parsing → structuring → verifying → ready

Use the poll script to wait for completion:

{baseDir}/scripts/shelv-poll-status.sh $SHELF_ID

If the shelf reaches failed, the error message and failed step are printed. You can retry:

curl -X POST "https://api.shelv.dev/v1/shelves/$SHELF_ID/retry" \
  -H "Authorization: Bearer $SHELV_API_KEY"

For review mode shelves, approve to finalize:

curl -X POST "https://api.shelv.dev/v1/shelves/$SHELF_ID/approve" \
  -H "Authorization: Bearer $SHELV_API_KEY"

Or regenerate the structure:

curl -X POST "https://api.shelv.dev/v1/shelves/$SHELF_ID/regenerate" \
  -H "Authorization: Bearer $SHELV_API_KEY"

Status-Gated Endpoint Matrix

Not all endpoints are available in every status. The archive, tree, and file endpoints require ready or review:

EndpointProcessingreviewreadyfailed
GET /v1/shelves/{id}YesYesYesYes
GET .../treeNoYesYesNo
GET .../files/*NoYesYesNo
GET .../archive-urlNoYesYesNo
POST .../approveNoYesNoNo
POST .../regenerateNoYesNoNo
POST .../retryNoNoNoYes

A 409 Conflict is returned if you call an endpoint outside its allowed statuses.

Rate Limits

ScopeLimit
Reads (GET)120 requests / min
Writes (POST/DELETE)20 requests / min
Shelf creation10 / hour

On 429 Too Many Requests, back off and retry after the indicated period.

Reference Pointers

For detailed API documentation, error codes, and lifecycle diagrams, see:

  • {baseDir}/references/api-reference.md — Full endpoint docs and response shapes
  • {baseDir}/references/shelf-lifecycle.md — Status flow, review mode, template behavior
  • {baseDir}/references/error-handling.md — Error codes, retry strategies

平台分布

OpenClaw

87.92%
按下载量换算5,545

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

未展示

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。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