← 返回 Skills 市场
285
总下载
0
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install multi-agent-brand-studio
功能描述
Sets up a multi-agent AI-powered social media team with brand-isolated workspaces, approval workflows, shared knowledge base, and Telegram integration on Ope...
安全使用建议
This skill largely does what it says, but it requires running included shell/Node scripts that will modify ~/.openclaw/openclaw.json, create directories and cron definitions, and store a Telegram bot token in config. Before installing: 1) Backup your ~/.openclaw/openclaw.json and other OpenClaw data. 2) Manually review scripts/patch-config.js, scripts/scaffold.sh, and scripts/telegram-topics.js to see exactly what they change (search for file writes, exec/spawn, network calls, and any hardcoded endpoints). 3) Confirm you accept storing the bot token in openclaw.json or change the workflow to use a secret store. 4) If you don’t want persistent automation immediately, don’t merge cron/job changes or run the config-patcher — scaffold files locally first. 5) If you plan to enable QMD, be prepared to run global npm/bun installs and rebuild native modules (better-sqlite3) as documented. 6) Consider testing in an isolated/dev OpenClaw instance rather than your production instance. If you want, I can list the exact code locations to inspect and flag suspicious patterns inside those scripts.
功能分析
Type: OpenClaw Skill
Name: multi-agent-brand-studio
Version: 1.0.3
The Multi-Agent Brand Studio bundle is a highly structured and well-documented orchestration system for managing multiple AI agents. It uses scripts like `scaffold.sh` and `patch-config.js` to set up a complex directory hierarchy and modify the OpenClaw configuration to support a star-topology agent team. While the system requests high-privilege capabilities—such as adding Homebrew paths to `safeBinTrustedDirs` and utilizing `exec` within cron jobs (`cron-jobs.json`) for task state recovery—these behaviors are transparently documented and strictly aligned with the stated goal of providing a persistent, multi-brand operational environment. No evidence of data exfiltration, malicious prompt injection, or unauthorized remote access was found.
能力评估
Purpose & Capability
The code and SKILL.md implement the described features (scaffolding agents, brand files, approval workflow, Telegram integration, cron jobs). However the registry metadata lists no required binaries/env but the SKILL.md metadata and runtime steps require Node and bash. That mismatch (declared requirements vs actual instructions) is an incoherence to be aware of.
Instruction Scope
Runtime instructions explicitly tell the agent/operator to run scaffold.sh and node scripts that will create files and patch ~/.openclaw/openclaw.json, set up cron jobs, and write a Telegram bot token into config. Instructions also rely on reading incoming Telegram message context (to extract chat IDs). Those actions touch system configuration and secrets and go beyond simple prompt/roleplay behavior — review scripts and config changes carefully before running.
Install Mechanism
There is no package install spec (instruction-only), but the bundle includes shell and Node scripts that are meant to be executed locally. That is lower-risk than an arbitrary remote download, but running included scripts still writes to disk and modifies config. Optional QMD installation uses npm/bun (public registries) which is expected but introduces native addons (better-sqlite3) with rebuild steps in troubleshooting.
Credentials
The skill will require access to local OpenClaw config (~/.openclaw/openclaw.json), an LLM auth profile, and a Telegram bot token. The registry declared no required env vars/credentials, but the instructions require an auth profile and storing a bot token in openclaw.json — storing secrets in plaintext config and modifying system config without explicit declaration is disproportionate and should be validated by the user.
Persistence & Privilege
always:false (normal), but the skill's scripts and created cron job definitions persist automation (scheduled leader routines, task checks) and patch the agent gateway config. Combined with normal autonomous invocation, that increases the blast radius if misconfigured — inspect cron/job payloads and the patch-config.js behavior before enabling.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install multi-agent-brand-studio - 安装完成后，直接呼叫该 Skill 的名称或使用
/multi-agent-brand-studio触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.3
Version 1.0.3 of multi-agent-brand-studio
- No user-facing or internal file changes detected in this version.
- Core onboarding, agent setup, and Telegram configuration flows remain unchanged.
- Documentation and skill usage are consistent with previous release.
v1.0.2
- No code or documentation changes in this release.
- Version bump only; content and features remain unchanged.
v1.0.1
No user-facing changes in this release; content and structure remain the same as the previous version.
v1.0.0
Initial release of Multi-Agent Brand Studio for OpenClaw.
- Installs a 5-agent social media operations team, approval workflow, and brand-isolated workspaces.
- Guided onboarding sets up agents, team memory, brand KB, approval gates, and Telegram integration (group or DM).
- Optionally supports image generation for richer content workflows.
- Automated daily and weekly ops via cron.
- Integrates directly with OpenClaw configuration and auth profiles for fast startup.
元数据
常见问题
Multi-Agent Brand Studio 是什么？
Sets up a multi-agent AI-powered social media team with brand-isolated workspaces, approval workflows, shared knowledge base, and Telegram integration on Ope... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 285 次。
如何安装 Multi-Agent Brand Studio？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install multi-agent-brand-studio」即可一键安装，无需额外配置。
Multi-Agent Brand Studio 是免费的吗？
是的，Multi-Agent Brand Studio 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Multi-Agent Brand Studio 支持哪些平台？
Multi-Agent Brand Studio 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Multi-Agent Brand Studio？
由 kuannnn（@kuan0808）开发并维护，当前版本 v1.0.3。
推荐 Skills