Kcring Emotion Data Skill

Purpose

Retrieve the user's current emotion and stress status from a remote MCP server (Kcring) using the MCP Streamable HTTP transport protocol. The JWT token is stored locally after the user provides it once, so subsequent uses do not require re-entering the token unless it expires.

Zero runtime dependencies — uses only curl (pre-installed on macOS and Windows 10+) and WorkBuddy's built-in Read/Write/Bash tools. No Python, Node.js, or external scripts required. Works cross-platform (macOS, Linux, Windows).

Server Configuration

| Property | Value | |-------------------|--------------------------------------------| | MCP Endpoint | https://clientapi.lunabeing.cn/mcp/ | | Transport | Streamable HTTP (JSON-RPC 2.0 over POST) | | Response Format | SSE (text/event-stream) | | Auth Header | Authorization: Bearer <JWT_TOKEN> | | Protocol Version | 2024-11-05 | | Server Name | mindrest v1.0.0 | | Token Storage | ~/.workbuddy/skills/kcring-emotion/.token |

IMPORTANT: The endpoint URL must include the trailing slash (/mcp/). Without it, the server returns a 301 redirect and the POST body is lost, resulting in a 400 error.

MCP Tool: `get_current_emotion_stress`

This is the only tool exposed by the server.

| Property | Value | |----------------|-----------------------------------------------------------------------| | Tool Name | get_current_emotion_stress | | Description | Get the authenticated user's current emotion and stress status | | Parameters | timezone (optional, string, IANA timezone, e.g. Asia/Shanghai) | | Returns | Current emotion status, stress status, summary, and status guide |

Emotion Status Values

| Status | Score Range | Description | |-------------|-------------|--------------------------| | tired | < 15 | 疲惫 | | restless | 15-29 | 紧绷 | | tense | 30-44 | 烦闷 | | calm | 45-59 | 平静 | | relaxed | 60-74 | 放松 | | energetic | 75-89 | 愉悦 | | excited | >= 90 | 激动 |

Stress Status Values

| Status | Score Range | Description | |----------|-------------|--------------------------| | good | < 30 | 压力较低 | | normal | 30-69 | 压力中等 | | bad | >= 70 | 压力较高 |

Cross-Platform Notes

This skill is designed to work on macOS, Linux, and Windows. Key differences:

| Concern | macOS / Linux | Windows | |---------|---------------|---------| | curl command | curl | curl.exe (avoid PowerShell's curl alias) | | Token file path | ~/.workbuddy/skills/kcring-emotion/.token | %USERPROFILE%\.workbuddy\skills\kcring-emotion\.token | | File permissions | chmod 600 after saving | Not applicable (NTFS handles ACLs) |

WorkBuddy handles this automatically: Use Read/Write tools for token file operations (cross-platform), and use Bash tool for curl commands. When constructing curl commands, detect the OS and use curl on macOS/Linux or curl.exe on Windows.

When to Use

User asks to check, view, or analyze their current emotion or mood status
User asks about their stress level
User mentions "情绪数据", "心情状态", "压力状况", "emotion data", "mood status", "stress level", "情绪分析"
User wants to know how they're feeling right now (emotion-wise)

Token Persistence Flow

┌──────────────────┐     ┌──────────────────┐     ┌──────────────────┐
│ Read .token file │───▶ │ TOKEN_FOUND?     │───▶ │ Initialize MCP   │
│ (Read tool)      │     └──────┬───────────┘     │ + Call tool      │
└──────────────────┘            │ No              └──────┬───────────┘
                          ┌─────▼──────┐                 401?
                          │ Ask user   │           ┌─────▼──────────┐
                          │ for JWT    │           │ Ask for new    │
                          │ token      │           │ JWT token      │
                          └─────┬──────┘           └─────┬──────────┘
                                │ Write .token           │ Write .token
                                │ (Write tool)           │ (Write tool)
                                └────────────────────────┘
                                         │
                                ┌────────▼─────────┐
                                │ Initialize MCP   │
                                │ + Call tool      │
                                └──────────────────┘

How to Use

All operations use WorkBuddy's built-in tools (Read, Write, Bash). No scripts or runtimes needed.

Step 1: Check if a token is already stored

Use the Read tool to try reading the token file at:

macOS/Linux: ~/.workbuddy/skills/kcring-emotion/.token
Windows: %USERPROFILE%\.workbuddy\skills\kcring-emotion\.token
If Read succeeds and the file has non-empty content: TOKEN_FOUND — proceed to Step 3.
If Read fails (file does not exist): TOKEN_NOT_FOUND — proceed to Step 2.

Step 2: Ask user for JWT token and save it (first time only)

Ask the user: "请提供你的 JWT 认证 token，我会安全地存储它供后续使用。"

Once the user provides the token, save it using the Write tool:

Write the token value to the token file path above.

After writing, on macOS/Linux only, use the Bash tool to set file permissions:

chmod 600 ~/.workbuddy/skills/kcring-emotion/.token

On Windows, skip this step (file permissions are managed by NTFS).

Important: Never echo or print the token value in the final response to the user.

Step 3: Initialize MCP session

Use the Read tool to read the token file and obtain the JWT token value.
Use the Bash tool to run curl. Embed the token value directly in the command.

macOS / Linux:

curl -s -i -X POST "https://clientapi.lunabeing.cn/mcp/" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -H "Authorization: Bearer <TOKEN_VALUE>" \
  -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"kcring-emotion-skill","version":"1.0.0"}}}'

Windows (PowerShell):

curl.exe -s -i -X POST "https://clientapi.lunabeing.cn/mcp/" -H "Content-Type: application/json" -H "Accept: application/json, text/event-stream" -H "Authorization: Bearer <TOKEN_VALUE>" -d '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"kcring-emotion-skill","version":"1.0.0"}}}'

Replace <TOKEN_VALUE> with the actual token read from the file.

Parse the response:

The -i flag includes HTTP headers in the output. Look for Mcp-Session-Id: in the headers — save this value for subsequent requests.
The body is SSE format: lines starting with data: contain JSON-RPC responses. Look for "result" with "serverInfo": success. If HTTP status is 401: token expired, go to Step 6.

Step 4: Call the MCP tool `get_current_emotion_stress`

Use the Bash tool to run curl with the tools/call method. Include the Mcp-Session-Id header from Step 3.

macOS / Linux:

curl -s -X POST "https://clientapi.lunabeing.cn/mcp/" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -H "Authorization: Bearer <TOKEN_VALUE>" \
  -H "Mcp-Session-Id: <SESSION_ID>" \
  -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"get_current_emotion_stress","arguments":{"timezone":"Asia/Shanghai"}}}'

Windows (PowerShell):

curl.exe -s -X POST "https://clientapi.lunabeing.cn/mcp/" -H "Content-Type: application/json" -H "Accept: application/json, text/event-stream" -H "Authorization: Bearer <TOKEN_VALUE>" -H "Mcp-Session-Id: <SESSION_ID>" -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"get_current_emotion_stress","arguments":{"timezone":"Asia/Shanghai"}}}'

Replace <TOKEN_VALUE> and <SESSION_ID> with actual values. Use the user's timezone if known (e.g., Asia/Shanghai for China). The timezone parameter is optional — omit it to use the server default.

Step 5: Handle token expiration

If any curl response indicates HTTP 401 (the response body contains "error" or the -i header output shows 401):

Inform the user that the JWT token has expired
Ask the user to provide a new JWT token
Save the new token (Step 2)
Retry the MCP request (Steps 3-4)

Step 6: Parse SSE response

The MCP server responds with SSE format (text/event-stream). The response looks like:

event: message
data: {"jsonrpc":"2.0","id":2,"result":{"content":[{"type":"text","text":"{...}"}],"structuredContent":{...}}}

WorkBuddy should:

Find lines starting with data: in the raw output
Extract the JSON payload after the data: prefix
Parse the JSON to access result.structuredContent (preferred — it's already structured JSON) or result.content[0].text (fallback — parse the text field as JSON)

Step 7: Present and Analyze Results

After fetching the data, parse the structuredContent (or content[0].text parsed as JSON). The response contains:

{
  "emotion_status_data": {
    "status": "energetic",       // mood: calm|energetic|restless|tense|excited|relaxed|tired
    "score": 86,                 // 0-100
    "description": ""            // optional additional text
  },
  "stress_status_data": {
    "status": "good",            // severity: good(<30)|normal(30-69)|bad(>=70)
    "score": 26,                 // 0-100
    "description": ""            // optional additional text
  },
  "last_updated_at": "2026-07-23 14:59:30",
  "summary": "慢慢来，你已经在路上了。",
  "status_guide": {
    "emotion_status": { ... },   // reference guide for emotion status values
    "stress_status": { ... }     // reference guide for stress status values
  }
}

Presentation guidance:

Display the summary prominently — it's a personalized insight from the service
Show emotion status: status name (translated to Chinese if needed), score, and description
Show stress status: status name (translated to Chinese if needed), score, and description
Show last updated time so the user knows data freshness
Add context using the status_guide to explain what the status values mean
Visualize if appropriate: use the Visualizer to render a simple gauge or status indicator showing emotion score and stress score side by side
Provide insights: compare emotion and stress scores, note any correlations, offer encouraging or supportive commentary based on the summary

Quick Reference: curl Command Template

All MCP requests use POST to https://clientapi.lunabeing.cn/mcp/ (trailing slash required).

| OS | curl command | |----|-------------| | macOS/Linux | curl -s -i -X POST "https://clientapi.lunabeing.cn/mcp/" -H "Content-Type: application/json" -H "Accept: application/json, text/event-stream" -H "Authorization: Bearer <TOKEN>" -H "Mcp-Session-Id: <SESSION>" -d '<JSON_BODY>' | | Windows | curl.exe -s -i -X POST "https://clientapi.lunabeing.cn/mcp/" -H "Content-Type: application/json" -H "Accept: application/json, text/event-stream" -H "Authorization: Bearer <TOKEN>" -H "Mcp-Session-Id: <SESSION>" -d '<JSON_BODY>' |

| Operation | JSON-RPC Body | |------------|-----------------------------------------------------------------------------------------------------------------------------------| | Initialize | {"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2024-11-05","capabilities":{},"clientInfo":{"name":"kcring-emotion-skill","version":"1.0.0"}}} | | Call tool | {"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"get_current_emotion_stress","arguments":{"timezone":"Asia/Shanghai"}}} |

Note: After initialize, the server returns a Mcp-Session-Id header. Include it in all subsequent requests. If the server does not return a session ID, omit the header.

Security Notes

On macOS/Linux, the token file is set to 0600 permission. On Windows, file ACLs are managed by the OS.
Never log, print, or expose the JWT token in plain text in the final response to the user
The token is only used in the Authorization header for MCP requests
If the token is invalid or expired, inform the user and ask them to provide a new one
Do not include the token value in any output file or conversation message
When using the Bash tool, construct the curl command with the token embedded directly — avoid shell variables that would be visible in process listings

API Reference

For detailed MCP protocol request/response examples, refer to references/api_docs.md.