Connections & Capabilities
Connects To
Capabilities
Quickstart
Config
{
"mcpServers": {
"crypto-sentiment-mcp": {
"command": "uv",
"args": ["--directory", "path/to/crypto-sentiment-mcp", "run", "main.py"],
"env": {
"SANTIMENT_API_KEY": "your_api_key_here"
}
}
}
}Exposed MCP Tools (5)
get_sentiment_balanceRetrieves the average sentiment balance for a specified cryptocurrency over a given period.
Read-only operation; retrieves data without modifying any system state.
get_social_volumeFetches the total number of social media mentions for a cryptocurrency.
Read-only operation; retrieves data without modifying any system state.
alert_social_shiftDetects significant spikes or drops in social volume for a cryptocurrency.
Read-only operation; analyzes data and reports anomalies without modifying any system state.
get_trending_wordsRetrieves the top trending words in cryptocurrency discussions.
Read-only operation; retrieves data without modifying any system state.
get_social_dominanceMeasures the percentage of crypto media discussions dominated by a specific cryptocurrency.
Read-only operation; retrieves data without modifying any system state.
Safety Assessment
The server primarily provides read-only access to cryptocurrency sentiment data, making it relatively safe. However, the risk of API key compromise and reliance on an external API should be considered. It is safe to use when the API key is securely managed and the limitations of sentiment analysis are understood. Risky scenarios involve using the data for automated trading without proper validation and risk management.
- Read-only access to cryptocurrency sentiment data.
- No direct write or delete operations on external systems.
- Relies on a third-party API (Santiment) for data retrieval.
- Requires an API key for authentication, limiting unauthorized access.
- Potential for information leakage if the API key is compromised.
- Dependence on the Santiment API introduces a single point of failure.
- Misinterpretation of sentiment data could lead to incorrect trading decisions by agents.
- No input sanitization is mentioned, which could lead to prompt injection if not handled correctly by the client.
