Token导航 LogoToken导航

Agent Skill · 研究检索

swarmclaw(swarmclaw 搜索)

swarmclaw 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。

swarmclaw 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,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 64,698。

总安装 294周安装 12GitHub Stars 156下载 94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:swarmclaw(swarmclaw 搜索)
来源仓库:https://github.com/swarmclawai/swarmclaw
仓库路径:skills/swarmclaw
安装命令:
npx skills add https://github.com/swarmclawai/swarmclaw --skill swarmclaw
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/swarmclawai/swarmclaw --skill swarmclaw

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。适合根据关键词或任务场景快速定位候选结果。可结合来源仓库和原始 README 核验具体用法。安装命令:npx skills add https://github.com/swarmclawai/swarmclaw --skill swarmclaw。安装前建议确认权限范围和维护状态。

来源仓库内路径:skills/swarmclaw

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/swarmclawai/swarmclaw --skill swarmclaw;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

swarmclaw 怎么安装?

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

swarmclaw 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。

swarmclaw 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。

swarmclaw 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。

swarmclaw 用的人多吗?

公开数据是累计安装 294 次,近一周安装 12 次,下载 94 次,GitHub 156 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。

swarmclaw 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

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 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。
  • 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
swarmclaw

swarmclaw 搜索 · 本页

294156接触敏感数据未明确通过

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

Overview

SwarmClaw is a self-hosted AI agent orchestration platform. This skill gives you CLI access to manage agents, tasks, chats, chatrooms, missions, schedules, memory, wallets, connectors, autonomy controls, webhooks, extensions, knowledge, and more across one or many SwarmClaw instances.

The CLI maps 1:1 to the SwarmClaw REST API. Every command follows the pattern:

swarmclaw <group> <action> [id] [--flags]

Install the CLI:

npm i -g @swarmclawai/swarmclaw

Setup

  1. Install: npm i -g @swarmclawai/swarmclaw
  2. Set env var: export SWARMCLAW_ACCESS_KEY=<your-key> (shown in terminal on first run)
  3. Default URL: http://localhost:3456 — override with SWARMCLAW_URL or --base-url
  4. Full docs: https://swarmclaw.ai/docs

Output Modes

Use machine-readable output when parsing results:

  • --json — structured JSON output (preferred for all API-mapped commands)
  • --raw — plain text output (legacy commands)

Filter results with --query key=value:

swarmclaw tasks list --json --query status=in_progress
swarmclaw memory list --json --query agentId=abc123 --query q=pricing
swarmclaw runs list --json --query sessionId=xyz --query limit=10

Core Commands

Agents

Manage your agent fleet — create, configure, monitor, and clone agents.

# List all agents
swarmclaw agents list --json

# Get full agent details
swarmclaw agents get <agentId> --json

# Create an agent (pass JSON body via --data)
swarmclaw agents create --data '{"name":"Research Bot","provider":"anthropic","model":"claude-sonnet-4-20250514"}' --json

# Update agent config
swarmclaw agents update <agentId> --data '{"name":"Updated Name","soul":"You are a research assistant"}' --json

# Get live agent status (running chats, current task, etc.)
swarmclaw agents status <agentId> --json

# Clone an agent with all its config
swarmclaw agents clone <agentId> --json

# Bulk update multiple agents
swarmclaw agents bulk-update --data '{"ids":["id1","id2"],"changes":{"provider":"openai"}}' --json

Other: delete, trash, restore, purge, thread

Tasks

Create, assign, and track work items on the task board.

# List all tasks (filterable by status, agent, project)
swarmclaw tasks list --json
swarmclaw tasks list --json --query status=in_progress --query agentId=<id>

# Create and assign a task
swarmclaw tasks create --data '{"title":"Analyse competitor pricing","description":"Research and compare competitor pricing strategies","agentId":"<agentId>"}' --json

# Get task details
swarmclaw tasks get <taskId> --json

# Update task status or fields
swarmclaw tasks update <taskId> --data '{"status":"completed"}' --json

# Get task board metrics (24h, 7d, or 30d)
swarmclaw tasks metrics --json --query range=7d

# Import GitHub issues as tasks
swarmclaw tasks import-github --data '{"repo":"owner/repo","labels":["bug"],"agentId":"<id>"}' --json

# Bulk update tasks
swarmclaw tasks bulk --data '{"ids":["id1","id2"],"changes":{"status":"todo"}}' --json

Other: delete, approve, claim, purge

Chats

Manage agent chat sessions — create, send messages, stream responses, and control execution.

Terminology: chats in the CLI = "sessions" in the SwarmClaw UI. Both refer to the same thing.
# List chats
swarmclaw chats list --json

# Create a new chat for an agent
swarmclaw chats create --data '{"agentId":"<agentId>"}' --json

# Send a message and stream the agent response (SSE)
swarmclaw chats chat <chatId> --data '{"message":"Give me a status update"}' --json

# Get chat message history
swarmclaw chats messages <chatId> --json

# Stop a running chat
swarmclaw chats stop <chatId> --json

# Clear chat history
swarmclaw chats clear <chatId> --json

# Append a user message without triggering a response
swarmclaw chats messages-send <chatId> --data '{"role":"user","content":"Note: budget is $5000"}' --json

Other: get, update, delete, delete-many, edit-resend, retry, queue, queue-add, queue-clear, heartbeat-disable-all, deploy, devserver, checkpoints

Chatrooms

Multi-agent collaborative spaces where multiple agents discuss and work together.

# List chatrooms
swarmclaw chatrooms list --json

# Create a chatroom
swarmclaw chatrooms create --data '{"name":"Strategy Room","description":"Multi-agent planning"}' --json

# Add agents to a chatroom
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<agent1>"}' --json
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<agent2>"}' --json

# Send a message — all member agents respond
swarmclaw chatrooms chat <chatroomId> --data '{"message":"What should our Q2 strategy be?"}' --json

# Remove an agent from the chatroom
swarmclaw chatrooms remove-member <chatroomId> --data '{"agentId":"<agentId>"}' --json

Other: get, update, delete, react, pin, moderate

Schedules

Automate recurring agent work on a cron schedule.

# List schedules
swarmclaw schedules list --json

# Create a schedule (cron syntax)
swarmclaw schedules create --data '{"name":"Daily Report","agentId":"<id>","cron":"0 9 * * *","message":"Generate the daily summary report"}' --json

# Update schedule
swarmclaw schedules update <scheduleId> --data '{"enabled":false}' --json

# Trigger a schedule immediately
swarmclaw schedules run <scheduleId> --json

# Delete a schedule
swarmclaw schedules delete <scheduleId> --json

Memory

Store and retrieve long-term agent memories for context persistence.

# List memories (searchable)
swarmclaw memory list --json
swarmclaw memory list --json --query q=pricing --query agentId=<id>

# Create a memory entry
swarmclaw memory create --data '{"agentId":"<id>","content":"Client prefers formal tone","type":"preference"}' --json

# Update a memory
swarmclaw memory update <memoryId> --data '{"content":"Updated content"}' --json

# Delete a memory
swarmclaw memory delete <memoryId> --json

# Analyse memory for dedup/prune candidates
swarmclaw memory maintenance --json

# Run the maintenance (dedupe/prune)
swarmclaw memory maintenance-run --json

Other: get, graph

Missions

Durable multi-step mission execution with control actions.

# List missions (filterable by status, phase, agent, project)
swarmclaw missions list --json
swarmclaw missions list --json --query status=active --query agentId=<id>

# Get mission detail
swarmclaw missions get <missionId> --json

# Get mission event timeline
swarmclaw missions events <missionId> --json

# Run a mission control action (resume, replan, retry_verification, wait, cancel)
swarmclaw missions action <missionId> --data '{"action":"resume"}' --json
swarmclaw missions action <missionId> --data '{"action":"cancel"}' --json

Search

Global cross-resource search across agents, tasks, chats, schedules, webhooks, and skills.

swarmclaw search query --json --query q=pricing

System Status

Lightweight health endpoint (v1.2.0+). Safe for external monitors and quick checks.

swarmclaw system-status get --json

More Commands

These groups are available but used less frequently by agents. Use swarmclaw <group> --help for full details.

GroupDescriptionKey commands
connectorsChat connectors (Discord, Slack, Telegram, etc.)list, get, create, update, delete, start, stop, repair, health, doctor
autonomySupervisor incidents, reflections, emergency stopincidents, reflections, estop, estop-set, guardian-restore
approvalsHuman-in-the-loop approval gateslist, resolve
webhooksInbound webhook triggers and delivery historylist, get, create, update, delete, trigger, history
walletsAgent wallets and transaction managementlist, get, create, update, delete, send, approve, transactions, balance-history
providersLLM provider configs and model overrideslist, get, create, update, delete, configs, ollama, models, models-set
gatewaysOpenClaw gateway profiles and health checkslist, create, update, delete, health
credentialsEncrypted provider credentialslist, get, create, delete
secretsEncrypted secret vaultlist, get, create, update, delete
notificationsIn-app notification centerlist, create, clear, mark-read, delete
extensionsExtension marketplace and configlist, set, install, marketplace, settings-get, settings-set, builtins
skillsReusable skill managementlist, get, create, update, delete, import
learned-skillsAgent-scoped learned skill reviewlist, promote, dismiss, delete, review-counts
knowledgeKnowledge base entries and document uploadlist, get, create, update, delete, upload
projectsProject grouping for agents and taskslist, get, create, update, delete
runsChat run queue and execution historylist, get, events
activityActivity feed eventslist (supports --query entityType=, --query action=)
daemonBackground daemon lifecyclestatus, start, stop, health-check
logsApplication logslist (supports --query lines=200, --query level=INFO), clear
setupDiagnostics and provider validationdoctor, check-provider
documentsDocument management and revisionslist, get, create, update, delete, revisions
uploadsUploaded artifact managementlist, get, delete, delete-many
clawhubBrowse and install ClawHub skillssearch, preview, install
openclawOpenClaw gateway control, deploy, sync (30+ subcommands)discover, gateway-status, deploy-*, remote-*, skills, sync

Rules

  1. Always verify an agent ID exists by running agents list before creating tasks for it.
  2. Never pass the access key as a CLI argument in visible output — rely on the SWARMCLAW_ACCESS_KEY env var.
  3. If the SwarmClaw instance is unreachable, run swarmclaw setup doctor --json and report the findings before retrying.
  4. When the user says "dispatch work", "get X agent to do Y", or "assign this to", use tasks create to assign the work.
  5. When the user asks "what's running", "agent status", or "what are my agents doing", combine output from agents list and chats list to give a full picture.
  6. For complex multi-step orchestration, create individual tasks rather than chaining commands.
  7. Prefer --json output mode for all commands. Use --raw only for legacy commands that don't support --json.
  8. Do not run commands that modify or delete agents without explicit user confirmation.
  9. Use system-status get for quick health checks — it's lightweight and safe for repeated polling.
  10. Use search query --query q=<term> to discover resources across agents, tasks, chats, schedules, and skills.
  11. The CLI group chats = "sessions" in the SwarmClaw UI. The legacy sessions alias is no longer available; always use chats.
  12. Use autonomy estop-set to engage emergency stop across all autonomous agents when safety action is needed.

Examples

Dispatch work to another agent

User says: "Get my research agent to analyse competitor pricing"

swarmclaw agents list --json
# Find the research agent's ID from the output
swarmclaw tasks create --data '{"title":"Analyse competitor pricing","description":"Research and compare competitor pricing strategies, identify gaps and opportunities","agentId":"<research-agent-id>"}' --json

Then confirm the task was created and the agent will pick it up.

Check fleet status

User says: "What are all my agents doing right now?"

swarmclaw agents list --json
swarmclaw chats list --json
swarmclaw tasks list --json --query status=in_progress

Then summarize which agents are idle, which have active chats, and any tasks in progress.

Run diagnostics

User says: "Something seems wrong with SwarmClaw"

swarmclaw system-status get --json
swarmclaw setup doctor --json

Check the health summary first (fast), then run full diagnostics if issues are found.

Multi-agent chatroom collaboration

User says: "Get my strategy and research agents to brainstorm Q2 plans together"

swarmclaw agents list --json
# Identify the strategy and research agent IDs
swarmclaw chatrooms create --data '{"name":"Q2 Strategy Brainstorm"}' --json
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<strategy-agent-id>"}' --json
swarmclaw chatrooms add-member <chatroomId> --data '{"agentId":"<research-agent-id>"}' --json
swarmclaw chatrooms chat <chatroomId> --data '{"message":"Brainstorm our Q2 growth strategy. Consider market trends, competitor moves, and our current strengths."}' --json

Schedule recurring work

User says: "Have my reporting agent send a daily summary every morning at 9am"

swarmclaw agents list --json
# Find the reporting agent's ID
swarmclaw schedules create --data '{"name":"Daily Morning Summary","agentId":"<reporting-agent-id>","cron":"0 9 * * *","message":"Generate and send the daily summary report covering key metrics, alerts, and action items from the last 24 hours"}' --json

Emergency stop all autonomous agents

User says: "Stop everything now!"

swarmclaw autonomy estop-set --data '{"engaged":true}' --json

Then confirm the emergency stop is engaged and all autonomous execution has halted.

Mission control

User says: "Check on the migration mission and resume it if it's paused"

swarmclaw missions list --json --query status=active
swarmclaw missions get <missionId> --json
swarmclaw missions events <missionId> --json
# If paused, resume it
swarmclaw missions action <missionId> --data '{"action":"resume"}' --json

Discovery

This skill covers the most-used commands. The full CLI has 50+ command groups. To explore:

swarmclaw --help                    # List all command groups
swarmclaw <group> --help            # List commands in a group
swarmclaw <group> <action> --help   # Show flags for a specific command

平台分布

Codex

34.01%
按下载量换算32

Claude

27.02%
按下载量换算25

Cursor

19.13%
按下载量换算18

Gemini CLI

9.97%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

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