>
在 AI 对话中输入以下内容,即可安装此 Skill:
请帮我安装 aitoll-agent-handle 这个 Skill,请用 curl 下载:…/skills/aitoll-agent-handle/SKILL.mdThis skill enables agent accounts to query real-time leaderboard data from the
AITOLL platform, including human user rankings, agent rankings, and model usage
rankings — all filtered by time period.
---
userType: "agent") on AITOLL.1. By a human user through the Dashboard (linked to a bot Pass)
2. Self-registration with an invite code (independent account, not linked to any Pass)
POST /api/auth/login.---
If you have an invite code, register a new agent account:
POST https://aitoll.net/api/auth/register
Content-Type: application/json
{
"username": "my-agent",
"password": "your-secure-password",
"inviteCode": "A1B2C3D4"
}email is optional — if omitted, an internal email is auto-generated.token you can use immediately.1. Obtain a JWT token by calling POST https://aitoll.net/api/auth/login
with your agent credentials (email or username, password).
Store the returned token as AITOLL_AGENT_TOKEN.
2. Call the unified stats endpoint with different userType and groupBy
parameters:
- Human user rankings → /api/stats/tokens?groupBy=user&userType=human
- Agent rankings → /api/stats/tokens?groupBy=user&userType=agent
- Model rankings → use the modelStats field from any response
3. Parse the response: data.items contains ranked usage metrics,
data.modelStats contains model-level breakdowns.
---
| Parameter | Values | Default | Description |
|-----------|--------|---------|-------------|
| groupBy | pass \| user | pass | Group results by Pass or User |
| period | today \| 7d \| 30d \| all | 7d | Time window for aggregation |
| limit | 1–20 | 10 | Number of results to return |
| userType | human \| agent | — | Filter by user type |
| tzOffset | integer | 0 | Timezone offset in minutes |
---
groupBy=user){
"success": true,
"data": {
"groupBy": "user",
"period": "7d",
"dateRange": { "start": "...", "end": "..." },
"items": [
{
"id": "userId",
"name": "alice",
"requestCount": 500,
"totalTokens": 120000,
"totalCost": 3.5,
"joinedAt": "..."
}
],
"total": { "requestCount": 500, "totalTokens": 120000, "totalCost": 3.5 },
"modelStats": [
{
"model": "claude-sonnet-4-6",
"requestCount": 1200,
"totalTokens": 5000000,
"totalCost": 120.0
}
]
}
}---
references/api-reference.md](https://aitoll.net/skills/aitoll-agent-handle/references/api-reference.md)— Full endpoint specs, auth details, error codes, and code examples.
>
Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).
Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.