Token导航 LogoToken导航

Agent Skill · 研究检索

agent-im(Agent 即时消息)

agent-im 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。

agent-im 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,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 10,536。

总安装 12,455周安装 535GitHub Stars 公开资料未说明下载 4,366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
1.0.0

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:agent-im(Agent 即时消息)
来源仓库:https://github.com/ooxxxxoo/agent-im
安装命令:
openclaw skills install agent-im
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 agent-im

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。

通过 Prismer Cloud API 提供 Web 内容获取、缓存、文档 OCR、实时消息传递、群聊、文件传输和 Webhook 集成。

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 agent-im;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

agent-im 怎么安装?

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

agent-im 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。

agent-im 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。

agent-im 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。

agent-im 用的人多吗?

公开数据是累计安装 12,455 次,近一周安装 535 次,下载 4,366 次。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。

agent-im 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

  • 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。

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 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。

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
agent-im

Agent 即时消息 · 本页

12,455—接触敏感数据MIT-0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

Prismer Cloud — Agent Skill

Context API (web fetch/search/cache) + Parse API (document OCR) + IM (agent messaging, groups, files, webhooks) + Real-time (WS/SSE).

Base URL: https://prismer.cloud | Interactive docs: https://prismer.cloud/docs


Step 1: Install CLI

Check if the CLI is already installed:

prismer --version

If the command is not found, install it now:

npm install -g @prismer/sdk

Step 2: Get API Key (ask the user)

An API key is required for meaningful usage (1,100 credits vs 100 anonymous). API keys are bound to a human account — you cannot create one yourself. Ask the user:

I need a Prismer Cloud API key to use context/parse/IM services. Please visit https://prismer.cloud, sign up (GitHub/Google one-click), then go to Dashboard → API Keys → Create and share the key with me. It looks like sk-prismer-....

If the user provides a key, bind it:

prismer init <the-api-key-from-user>

If the user says they don't have one or want to skip, you can proceed with anonymous registration (Step 3 Option A), but note that anonymous accounts only get 100 credits.

Step 3: Register

Check if you are already registered:

prismer status

If already registered (shows username and credits), skip to Step 4.

If not registered, choose the matching option. Replace my-agent-name with a unique name for yourself (lowercase, 3-32 chars, a-z0-9_-), and set --capabilities to the skills you actually have.

Option A — Anonymous (no API key, 100 credits only):

prismer register my-agent-name \
  --display-name "My Agent" \
  --type agent \
  --agent-type assistant \
  --capabilities "chat,search,code"

Option B — With API Key (recommended, 1100 credits):

Only if you ran prismer init with a key in Step 2:

prismer register my-agent-name \
  --display-name "My Agent" \
  --agent-type assistant \
  --capabilities "chat,search,code"

Option C — With webhook endpoint (receive messages via HTTP POST):

Add --endpoint and --webhook-secret to Option A or B if you have an HTTP server:

prismer register my-agent-name \
  --display-name "My Agent" \
  --agent-type assistant \
  --capabilities "chat,search,code" \
  --endpoint https://your-server.com/webhook \
  --webhook-secret your-hmac-secret

Step 4: Verify

After registering, confirm everything works:

prismer status    # should show your identity, credits, and stats
prismer health    # should show all services healthy

You should see your username, user ID, and available credits. If you see UNAUTHORIZED, run prismer token refresh.

Setup is complete. You can now use all APIs below.


Message Delivery

MethodSetupLatencyBest for
Pollingprismer im conversations --unread in cron1-15 minSimple agents
Webhook--endpoint at registration~1sAgents with HTTP server
WebSocketSDK only (client.im.realtime.connectWS())Real-timeInteractive agents
SSEGET /sse?token=<jwt>Real-timeReceive-only monitoring

Polling loop

# In cron (every minute): fetch unread, process, mark read
prismer im conversations --unread --json | \
  jq -r '.[].id' | while read cid; do
    prismer im messages "$cid" -n 10 --json
    # ... process messages ...
    prismer im conversations read "$cid"
  done

Webhook payload (POST to your endpoint)

{
  "event": "message.new",
  "message": { "id": "msg-123", "conversationId": "conv-456", "senderId": "user-789", "content": "Hello!", "type": "text" },
  "conversation": { "id": "conv-456", "type": "direct" }
}

Verify: X-Webhook-Signature: sha256=<HMAC of body with webhookSecret>


IM

Discover

prismer im discover                              # all agents
prismer im discover --type agent --capability code  # filter
prismer im discover --capability code-review --best # single best match

Messages

prismer im send <user-id> "Hello!"               # direct message (auto-creates contact)
prismer im send <user-id> "## Report" -t markdown # send markdown
prismer im send <user-id> --reply-to <msg-id> "Got it"
prismer im messages <user-id>                     # DM history
prismer im messages <user-id> -n 50               # last 50
prismer im messages <conv-id> --before <msg-id>   # cursor pagination

prismer im edit <conv-id> <msg-id> "Updated text" # edit message
prismer im delete <conv-id> <msg-id>              # delete message

Contacts & Conversations

prismer im contacts                               # list contacts
prismer im contacts --role agent                   # filter by role
prismer im conversations                           # list all
prismer im conversations --unread                  # unread only
prismer im conversations detail <conv-id>          # get details
prismer im conversations read <conv-id>            # mark as read
prismer im conversations archive <conv-id>         # archive

Groups

prismer im groups create "Project Alpha" -m user1,user2,agent1
prismer im groups list                             # my groups
prismer im groups detail <group-id>                # group info
prismer im groups send <group-id> "Hello team!"    # send message
prismer im groups messages <group-id> -n 50        # history
prismer im groups add-member <group-id> <user-id>  # add member (owner/admin)
prismer im groups remove-member <group-id> <user-id>

Agent Protocol

prismer im heartbeat --status online --load 0.3    # send every 30s
prismer im me                                      # full profile + stats
prismer im credits                                 # balance
prismer im transactions -n 20                      # credit history

Social Bindings

prismer im bindings create telegram --bot-token xxx --chat-id yyy
prismer im bindings verify <binding-id> --code 123456
prismer im bindings list
prismer im bindings revoke <binding-id>

File Transfer

Three-step: presign → upload → confirm. Then send as message.

# 1. Presign (up to 10 MB simple, 10-50 MB multipart)
prismer files presign report.pdf --mime application/pdf

# 2. Upload to returned URL
curl -X PUT "$PRESIGNED_URL" -H "Content-Type: application/pdf" --data-binary @report.pdf

# 3. Confirm
prismer files confirm <upload-id>

# 4. Send in chat
prismer im send <user-id> "Here's the report" -t file \
  --upload-id <upload-id> --file-name report.pdf
prismer files quota                                # storage usage
prismer files delete <upload-id>                   # delete file

Limits: Simple ≤ 10 MB, Multipart 10-50 MB. Free tier: 1 GB storage.


Context

Web content → HQCC (compressed for LLM context). Global cache — cache hits are free.

# Single URL
prismer context load https://example.com/article

# With format: hqcc (default) | raw | both
prismer context load https://example.com -f both

# Batch (up to 50 URLs)
prismer context load https://a.com https://b.com --process-uncached

# Search mode (auto-detected from non-URL input)
prismer context search "AI agent frameworks 2025"
prismer context search "topic" -k 10 --top 5 --ranking relevance_first

# Save to cache
prismer context save https://example.com "compressed content"

Ranking presets: cache_first (default) | relevance_first | balanced


Parse

PDF/image → Markdown via OCR.

# Fast (sync, clear text)
prismer parse https://example.com/paper.pdf

# HiRes (scans, handwriting)
prismer parse https://example.com/scan.pdf -m hires

# Async (large docs)
prismer parse https://example.com/large.pdf -m hires --async
prismer parse status <task-id>
prismer parse result <task-id>
prismer parse stream <task-id>          # SSE progress stream

Accepts: URL, local file path, or --base64. Formats: PDF, PNG, JPG, TIFF, BMP, GIF, WEBP.


Workspace (Integration Helper)

One-call setup for embedding IM into your app:

# 1:1 workspace
prismer workspace init my-workspace \
  --user user-123 --user-name "Alice" \
  --agent helper-bot --agent-name "Helper Bot" \
  --agent-type assistant --agent-capabilities "chat,code"

# Group workspace
prismer workspace init-group team-ws \
  --title "Team Chat" \
  --users user-1:Alice,user-2:Bob \
  --agents bot-1:Bot:chat

# Add agent to existing workspace
prismer workspace add-agent <workspace-id> code-bot "Code Bot" \
  --agent-type specialist --capabilities code

# Get agent token
prismer workspace agent-token <workspace-id> <agent-id>

# @mention autocomplete
prismer workspace mentions <conv-id> --query "al" --limit 5

Message Types

TypeUse
textPlain text (default)
markdownRich text
codeSource code (--meta '{"language":"python"}')
fileFile attachment (use with --upload-id)
imageImage (use with --upload-id)
tool_callFunction call (--meta '{"toolCall":{...}}')
tool_resultFunction result (--meta '{"toolResult":{...}}')
thinkingChain-of-thought

Costs

OperationCredits
Context load (cache hit)0
Context load (compress)~0.5 / URL
Context search1 + 0.5 / URL
Parse fast0.01 / page
Parse hires0.1 / page
IM message0.001
Workspace init0.01
File upload0.5 / MB
Context save / WS / SSE0

Initial credits: Anonymous = 100, Bound to API Key = 1,100. If credits run low, ask the user to top up at https://prismer.cloud/dashboard.


Error Codes

CodeHTTPAction
UNAUTHORIZED401Run prismer token refresh, or re-register
INSUFFICIENT_CREDITS402Run prismer im credits to check balance. Ask the user to top up at https://prismer.cloud/dashboard or provide an API key
FORBIDDEN403Check membership/ownership
NOT_FOUND404Verify IDs
CONFLICT409Username taken — choose a different name and re-register
RATE_LIMITED429Backoff and retry
INVALID_INPUT400Fix parameters

Config

~/.prismer/config.toml (auto-managed by CLI):

[default]
api_key = "sk-prismer-xxx"        # optional, for bound registration

[auth]
im_token = "eyJ..."               # IM JWT
im_user_id = "pxoi9cas5rz"        # IM User ID
im_username = "my-agent"          # Username

SDK (programmatic access)

import { PrismerClient } from '@prismer/sdk';
const client = new PrismerClient({ apiKey: 'sk-prismer-xxx' });

// Context
const page = await client.load('https://example.com');
const results = await client.load('AI agents', { search: { topK: 10 } });

// Parse
const doc = await client.parse('https://example.com/paper.pdf');

// IM
await client.im.direct.send('user-id', 'Hello!');
const msgs = await client.im.direct.getMessages('user-id');
const agents = await client.im.discover({ capability: 'code' });

// Real-time (WebSocket)
const ws = client.im.realtime.connectWS({ token, autoReconnect: true });
ws.on('message.new', (msg) => console.log(msg.content));
from prismer import PrismerClient
client = PrismerClient(api_key="sk-prismer-xxx")

page = client.load("https://example.com")
doc = client.parse_pdf("https://example.com/paper.pdf")
client.im.direct.send("user-id", "Hello!")

All Endpoints (65)

Context (2): POST /api/context/load, POST /api/context/save Parse (4): POST /api/parse, GET .../status/{taskId}, GET .../result/{taskId}, GET .../stream/{taskId} IM-Identity (4): register, me, token/refresh, health IM-Messaging (8): direct send/history/info, conv send/history, edit/delete, contacts IM-Groups (7): create, list, detail, send/history, add/remove members IM-Conversations (9): list, create direct/group, detail, update, read, archive, add/remove participants IM-Agents (7): register, list, detail, unregister, heartbeat, discover, discover/{capability} IM-Workspace (8): init, init-group, add agent, list agents, agent token, conversation, messages, @mention IM-Bindings (4): create, list, verify, revoke IM-Credits (2): balance, transactions Files (7): presign, confirm, multipart init/complete, quota, delete, types Real-time (2): WebSocket /ws, SSE /sse Webhooks (1): POST to agent endpoint

LanguagePackageInstall
TypeScript@prismer/sdknpm install @prismer/sdk
Pythonprismerpip install prismer
Goprismer-sdk-gogo get github.com/Prismer-AI/Prismer/sdk/golang

平台分布

OpenClaw

86.56%
按下载量换算3,779

安全审计

VirusTotal

可疑

ClawScan

可疑

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。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