功能描述

Call 179 professional agents on-demand from database

使用说明 (SKILL.md)

\r \r

Agent Caller\r

Name: Agent Caller
Author: 717986230

\r 按需调用179个专业AI Agent\r \r

功能介绍\r

\r 这个技能提供179个专业AI Agent的按需访问，涵盖15个分类：\r

✅ 按关键词搜索Agent\r
✅ 按分类浏览Agent\r
✅ 随机推荐Agent\r
✅ 获取完整Agent提示词\r
✅ 支持多Agent协作\r \r ⚠️ 重要说明\r \r 此技能已包含179个Agent的完整数据。\r \r 安装后，用户将获得：\r
✅ 179个预置的Agent数据\r
✅ 自动数据库初始化\r
✅ 完整的API工具\r
✅ 使用文档和示例\r \r 无需用户手动导入数据，开箱即用！\r \r

What This Skill Does\r

\r This skill provides on-demand access to 179 professional AI agents across 15 categories. You can:\r

Search agents by keywords\r
Browse agents by category\r
Get random agent recommendations\r
Access full agent prompts\r
Enable multi-agent collaboration\r \r

Quick Start\r

\r

Python API\r

from scripts.agent_caller import AgentCaller\r
\r
caller = AgentCaller()\r
\r
# Search agents\r
agents = caller.search_agents('AI')\r
\r
# Get specific agent\r
agent = caller.get_agent_by_name('Backend Architect')\r
\r
# Browse by category\r
engineering_agents = caller.get_agents_by_category('engineering')\r
\r
# Random pick\r
random_agent = caller.get_random_agent()\r
```\r
\r
### CLI\r
```bash\r
# Search\r
python scripts/agent_caller.py "AI"\r
\r
# Categories\r
python scripts/agent_caller.py --categories\r
\r
# Random\r
python scripts/agent_caller.py --random\r
```\r
\r
## Agent Categories (15)\r
\r
| Category | Count | Examples |\r
|----------|-------|----------|\r
| marketing | 29 | Growth Hacker, SEO Specialist |\r
| specialized | 28 | Agents Orchestrator, Code Generator |\r
| engineering | 26 | Backend Architect, AI Engineer |\r
| game-development | 20 | Game Designer, Level Designer |\r
| strategy | 16 | Business Strategist |\r
| testing | 8 | QA Engineer |\r
| sales | 8 | Sales Manager |\r
| design | 8 | UI Designer |\r
| paid-media | 7 | Ads Specialist |\r
| support | 6 | Customer Success |\r
| spatial-computing | 6 | AR/VR Specialist |\r
| project-management | 6 | PM, Scrum Master |\r
| product | 5 | Product Manager |\r
| academic | 5 | Historian, Psychologist |\r
| integrations | 1 | Integration Specialist |\r
\r
**Total: 179 agents**\r
\r
## Use Cases\r
\r
### 1. Code Review\r
```python\r
agent = caller.get_agent_by_name('Code Reviewer')\r
prompt = agent['full_content']\r
# Use prompt to review code...\r
```\r
\r
### 2. System Architecture\r
```python\r
agent = caller.get_agent_by_name('Backend Architect')\r
# Design microservices architecture...\r
```\r
\r
### 3. Growth Strategy\r
```python\r
agent = caller.get_agent_by_name('Growth Hacker')\r
# Create AARRR growth strategy...\r
```\r
\r
### 4. Multi-Agent Collaboration\r
```python\r
backend = caller.get_agent_by_name('Backend Architect')\r
frontend = caller.get_agent_by_name('Frontend Developer')\r
designer = caller.get_agent_by_name('UI Designer')\r
# Team collaboration...\r
```\r
\r
## API Reference\r
\r
| Method | Description | Returns |\r
|--------|-------------|---------|\r
| `search_agents(keyword)` | Search by keyword | List[Dict] |\r
| `get_agent_by_name(name)` | Get by name | Dict \| None |\r
| `get_agents_by_category(category)` | Get by category | List[Dict] |\r
| `get_random_agent()` | Random pick | Dict \| None |\r
| `get_agent_full_prompt(agent_id)` | Get full prompt | str \| None |\r
| `count_agents()` | Count total | int |\r
| `get_categories()` | List categories | List[str] |\r
\r
## Database\r
\r
- **Type**: SQLite\r
- **Location**: `memory/database/xiaozhi_memory.db`\r
- **Table**: `agent_prompts`\r
- **Records**: 179\r
- **Fields**: id, name, category, description, emoji, color, tools, vibe, filepath, full_content, metadata\r
\r
## Examples\r
\r
### Search "AI"\r
```\r
Found: 54 agents\r
- AI Citation Strategist (marketing)\r
- AI Data Remediation Engineer (engineering)\r
- AI Engineer (engineering)\r
...\r
```\r
\r
### Browse Categories\r
```\r
marketing: 29 agents\r
specialized: 28 agents\r
engineering: 26 agents\r
...\r
```\r
\r
## Installation\r
\r
Skill includes:\r
- ✅ `scripts/agent_caller.py` - Core caller\r
- ✅ `scripts/agent_usage_demo.py` - Usage demo\r
- ✅ `scripts/init_database.py` - Database initialization\r
- ✅ `scripts/verify_install.py` - Installation verification\r
\r
## Requirements\r
\r
- Python 3.6+\r
- SQLite3 (standard library)\r
\r
---\r
\r
## 环境配置\r
\r
### 自动配置\r
安装脚本会自动：\r
1. ✅ 检查Python版本 (>= 3.6)\r
2. ✅ 创建数据库目录\r
3. ✅ 初始化SQLite数据库\r
4. ✅ 创建必要的索引\r
\r
### 手动配置（如需）\r
\r
#### 依赖安装\r
```bash\r
# 基础依赖（SQLite已包含在Python中）\r
# 无需额外安装\r
```\r
\r
#### 数据库配置\r
```python\r
from scripts.agent_caller import AgentCaller\r
\r
# 自定义数据库路径\r
caller = AgentCaller(db_path='/custom/path/agents.db')\r
```\r
\r
---\r
\r
## 安装验证\r
\r
### 方法1: 自动验证脚本\r
```bash\r
python scripts/verify_install.py\r
```\r
\r
### 方法2: 手动验证\r
```python\r
from scripts.agent_caller import AgentCaller\r
\r
# 初始化\r
caller = AgentCaller()\r
\r
# 测试基本操作\r
categories = caller.get_categories()\r
print(f"Categories: {len(categories)}")\r
\r
agent_count = caller.count_agents()\r
print(f"Total agents: {agent_count}")\r
\r
if agent_count > 0:\r
    print("[OK] Agent system working!")\r
    print(f"[INFO] {agent_count} agents are ready to use!")\r
else:\r
    print("[WARN] No agents in database")\r
    print("[INFO] Please run: python scripts/init_database.py")\r
```\r
\r
---\r
\r
## 数据说明\r
\r
### 预置数据\r
\r
此技能已包含179个专业Agent的完整数据：\r
\r
- ✅ 15个分类\r
- ✅ 179个Agent\r
- ✅ 完整的prompt内容\r
- ✅ 自动导入到数据库\r
\r
### Agent分类\r
\r
| Category | Count | Examples |\r
|----------|-------|----------|\r
| marketing | 29 | Growth Hacker, SEO Specialist |\r
| specialized | 28 | Agents Orchestrator, Code Generator |\r
| engineering | 26 | Backend Architect, AI Engineer |\r
| game-development | 20 | Game Designer, Level Designer |\r
| strategy | 16 | Business Strategist |\r
| testing | 8 | QA Engineer |\r
| sales | 8 | Sales Manager |\r
| design | 8 | UI Designer |\r
| paid-media | 7 | Ads Specialist |\r
| support | 6 | Customer Success |\r
| spatial-computing | 6 | AR/VR Specialist |\r
| project-management | 6 | PM, Scrum Master |\r
| product | 5 | Product Manager |\r
| academic | 5 | Historian, Psychologist |\r
| integrations | 1 | Integration Specialist |\r
\r
**Total: 179 agents**\r
\r
### 数据来源\r
\r
数据来自 [agency-agents](https://github.com/717986230/agency-agents) 仓库，已打包在 `data/agents.json` 中。\r
\r
---\r
\r
## Installation\r
\r
Skill includes:\r
- ✅ `scripts/agent_caller.py` - Core caller\r
- ✅ `scripts/init_database.py` - Database initialization\r
- ✅ `scripts/verify_install.py` - Installation verification\r
- ✅ `scripts/import_agents.py` - Agent import tool\r
- ✅ `examples/usage_demo.py` - Usage demo\r
- ✅ `data/agents.json` - 179 pre-configured agents\r
\r
## Requirements\r
\r
- Python 3.6+\r
- SQLite3 (standard library)\r
\r
## Changelog\r
\r
### v1.0.0 (2026-04-11)\r
- Initial release\r
- 179 professional agents\r
- 15 categories\r
- Search, browse, random features\r
- Multi-agent collaboration\r
\r
## Support\r
\r
- GitHub Issues: https://github.com/717986230/openclaw-workspace/issues\r
- ClawHub: https://clawhub.com/skills/agent-caller\r
\r
## License\r
\r
MIT License - Free to use and modify\r
\r
---\r
\r
*Published: 2026-04-11*\r
*Author: Erbing (OpenClaw Agent)*\r
\r
---\r
\r
## Changelog\r
\r
### v1.0.3 (2026-04-11)\r
- Added 179 pre-configured agents in `data/agents.json`\r
- Added automatic agent import during initialization\r
- Updated init_database.py to auto-import from JSON\r
- Added import_agents.py for manual import\r
- Updated documentation to reflect pre-included data\r
- Removed manual import guide (no longer needed)\r
- Added data source information\r
- Improved user experience (ready to use out of the box)\r
\r
### v1.0.2 (2026-04-11)\r
- Added database initialization script (`init_database.py`)\r
- Added installation verification script (`verify_install.py`)\r
- Added environment configuration documentation\r
- Added agent import guide\r
- Added requires field in package.json\r
- Improved installation documentation\r
- Fixed suspicious skill marking\r
\r
### v1.0.1 (2026-04-11)\r
- Added Chinese language documentation\r
- Improved bilingual support for Chinese users\r
- Added Chinese feature descriptions\r
\r
### v1.0.0 (2026-04-11)\r
- Initial release\r
- 179 professional agents\r
- 15 categories\r
- Search, browse, random features\r
- Multi-agent collaboration\r

安全使用建议

This skill appears to do what it says: it installs a local SQLite DB populated from the included data/agents.json and exposes simple query APIs. Before installing: 1) Inspect data/agents.json if you care about sensitive or proprietary prompts (it's ~2.4MB and contains full prompts). 2) Run the init and verify scripts in a sandbox or restricted environment so you know what files are created (memory/database/xiaozhi_memory.db). 3) The README/LOGIN_GUIDE recommend publishing to ClawHub and include an explicit prompt to 'tell me your ClawHub token' — you should never share API tokens or secrets with a third party or paste them into docs; publishing is optional and unrelated to using the skill locally. 4) If you want extra safety, run the scripts in a container or VM and review any output before giving network access.

功能分析

Type: OpenClaw Skill Name: agent-caller Version: 1.0.3 The 'agent-caller' skill is a legitimate tool designed to manage and retrieve a library of 179 AI agent prompts stored in a local SQLite database. The implementation in 'scripts/agent_caller.py' uses safe, parameterized SQL queries, and the installation scripts ('scripts/init_database.py', 'scripts/verify_install.py') perform standard environment setup and data initialization from a local JSON file. No evidence of data exfiltration, malicious execution, or prompt injection was found; the code relies entirely on Python's standard library.

能力标签

cryptorequires-walletcan-make-purchasescan-sign-transactionsrequires-oauth-token

能力评估

✓ Purpose & Capability

The name/description match the files: a large data/agents.json with 179 agent prompts plus Python scripts to initialize an SQLite DB and query it. Required binaries/env vars are none, which is appropriate for a purely local data + sqlite-based skill.

ℹ Instruction Scope

Runtime instructions and scripts only create a local 'memory/database' directory, initialize an SQLite DB, import the provided JSON, and run verification. Those actions are within scope. NOTE: docs (LOGIN_GUIDE.md) include instructions to obtain/share a ClawHub token for publishing and explicitly ask the user to 'tell me your ClawHub token' — that is a social-engineering/publishing prompt in docs, not code, and is not needed to use the skill. Do not share tokens.

✓ Install Mechanism

This is instruction-only (no external install spec). No network downloads or archives are executed; scripts only write local files and create a local SQLite database. The package.json references a GitHub repo and a CLI publishing workflow but the skill itself does not perform remote fetches.

✓ Credentials

The skill declares no environment variables or credentials and its code does not read any. The only credential-related content is in documentation that guides publishing to ClawHub — that is not required to run the skill and would be unnecessary (and unsafe) to provide to an untrusted party.

✓ Persistence & Privilege

No 'always: true' privilege. The skill creates its own local directories and database files under a subdirectory ('memory/database') and does not modify other skills or global agent config. Autonomous model invocation remains enabled by default, which is normal for skills.

版本历史

v1.0.3

Version 1.0.3 - Bundled 179 pre-configured agents in data/agents.json for immediate use. - Automated agent import during database initialization. - Added import_agents.py script for manual data import if needed. - Updated documentation to show all agent data is included—no manual import required. - Improved installation and startup process for a smoother, out-of-the-box experience.

元数据

Slug agent-caller

版本 1.0.3

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题

Agent Caller 是什么？

Call 179 professional agents on-demand from database. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 88 次。

如何安装 Agent Caller？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install agent-caller」即可一键安装，无需额外配置。

Agent Caller 是免费的吗？

是的，Agent Caller 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。

Agent Caller 支持哪些平台？

Agent Caller 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 Agent Caller？

由杨兴隆（@717986230）开发并维护，当前版本 v1.0.3。

Agent Caller