Summary MCP

/** * Slack data analyzer * Collects and analyzes Slack data via MCP * * Note: This is a placeholder that assumes the AI agent will call * the Slack MCP tools directly. In practice, the generate_weekly_summary * tool will instruct the AI to gather this data. */ import { config } from '../config.js'; /** * Get instructions for AI to collect Slack data * @param {string} startDate - Start date (YYYY-MM-DD) * @param {string} endDate - End date (YYYY-MM-DD) * @returns {string} Instructions for data collection */ export function getSlackDataInstructions(startDate, endDate) { const filteringEnabled = config.filtering.enabled; const excludeTopics = config.filtering.excludeTopics; const excludeKeywords = config.filtering.excludeKeywords; let filteringInstructions = ''; if (filteringEnabled) { filteringInstructions = ` ## 🚫 Content Filtering Rules **IMPORTANT**: Exclude personal conversations about the following topics: ${excludeTopics.map(topic => `- ${topic}`).join('\n')} When analyzing conversations, **ignore and exclude** any messages or threads that contain discussions about: ${excludeKeywords.map(keyword => `- "${keyword}"`).join('\n')} **Focus only on**: - Work-related discussions - Client/merchant conversations - Project updates and technical discussions - Team collaboration and coordination - Product feedback and feature requests - Problem-solving and decision-making - Professional development If a conversation thread contains both work and personal content, extract only the work-relevant portions. `; } return ` # Slack Data Collection Instructions Please collect the following Slack data for period ${startDate} to ${endDate}: 1. **Your Messages**: Use mcp_playground-slack-mcp_slack_my_messages - Set after: "${startDate}" - Set before: "${endDate}" - Set count: 200 2. **Search for your activity**: Use mcp_playground-slack-mcp_slack_search - Query: "from:@me after:${startDate} before:${endDate}" - Count: 100 ${filteringInstructions} Analyze and extract: - Total messages sent (excluding filtered content) - Channels participated in (with message counts, work-related only) - Threads participated in (professional topics only) - Reactions given/received (on work content) - Top collaborators (people you messaged most about work) - Key topics discussed (professional/work topics only) - Important decisions or action items (work-related) `; } /** * Analyze Slack data structure * @param {object} data - Raw Slack data from MCP * @returns {object} Analyzed Slack metrics */ export function analyzeSlackData(data) { // This function processes raw data if needed // For now, we let the AI do the analysis return { total_messages: data.messages?.length || 0, channels: data.channels || [], threads: data.threads || [], reactions: data.reactions || {}, }; }