Attio MCP Server

# Universal Tools Quick Reference: Copy-Paste Prompts _Validated prompts that work perfectly with the Attio MCP Server's universal tools architecture. Every prompt tested for compatibility with the current 14-tool system._ ## 🎯 Tool Architecture Overview The Attio MCP Server uses **14 universal tools** (68% reduction from 40+ tools) that handle all CRM operations through `resource_type` parameters: | **Universal Tool** | **Purpose** | **Resource Types** | | -------------------------------- | ---------------------------------- | ----------------------------------- | | `records.search` | Basic search with filters | companies, people, tasks, deals | | `records.search_advanced` | Multi-condition complex queries | companies, people, tasks, deals | | `records.get_details` | Complete record information | companies, people, tasks, deals | | `create-record` | Create new records | companies, people, tasks, deals | | `update-record` | Modify existing records | companies, people, tasks, deals | | `delete-record` | Remove records | companies, people, tasks, deals | | `records.get_attributes` | Available fields for resource type | companies, people, tasks, deals | | `records.discover_attributes` | Dynamic field discovery | companies, people, tasks, deals | | `records.get_info` | Structured info by type | companies, people | | `records.search_by_relationship` | Find connected records | companies ↔ people | | `records.search_by_content` | Text-based matching | notes, activities, content | | `records.search_by_timeframe` | Date-based filtering | created, modified, last_interaction | | `records.batch` | Bulk processing | create, update, delete, search | | `records.search_batch` | Multi-criteria batch search | All resource types | **Plus 11 Lists Tools**: Complete pipeline and list management functionality --- ## 🚀 Validated Copy-Paste Prompts ### 🔍 Basic Search Operations #### **Find Companies (Universal)** ``` Use records.search with resource_type="companies" to find: - Technology companies with 50+ employees - Show: name, industry, employee_count, website - Limit to 15 results for optimal performance ``` #### **Advanced Company Search** ``` Use records.search_advanced for companies where: - Industry contains "Technology" OR "Software" - Employee count greater than 100 - Created in last 60 days - Resource type: companies Format as table with company name, size, industry, date added ``` #### **Find People by Role** ``` Search for people using records.search_advanced: - Resource type: people - Job title contains "CEO" OR "Founder" OR "VP" - Company size over 50 employees - Show: name, title, company, email Limit to 20 results to avoid truncation ``` ### 📊 Pipeline Management #### **Sales Pipeline Overview** ``` Get my sales pipeline status using these steps: 1. Use get-lists to see all available pipelines 2. For each pipeline, use get-list-entries to show current deals 3. Use advanced-filter-list-entries for deals over $5,000 4. Group by stage and calculate totals Limit each stage to top 10 deals by value ``` #### **Move Deal Through Pipeline** ``` Update pipeline position: 1. Find the deal using records.search with resource_type="deals" 2. Use update-list-entry to change stage from "Qualified" to "Proposal" 3. Use add-record-to-list if moving between different pipelines 4. Confirm the update with get-list-entries Include deal value and expected close date in the update ``` ### ✅ Task Management (Addresses Issue #472) #### **Create Tasks (Avoiding "Unnamed" Issue)** ``` Create follow-up tasks using create-record: - Resource type: tasks - Content: "Follow up on demo with [Company Name] - discuss pricing" - Title: "Demo Follow-up: [Company Name]" - Due date: [specific date] - Assigned to: [team member] IMPORTANT: Always include both 'content' and 'title' fields to avoid "Unnamed" display ``` #### **Review Tasks (Workaround for Issue #472)** ``` Get task overview using these methods: 1. Use records.search_by_timeframe for tasks created in last 7 days 2. Use records.get_details for each task to get complete info 3. Show: content, title, due_date, assignee, linked_company Avoid generic "show all tasks" - use specific timeframes or filters ``` ### 🔗 Relationship Discovery #### **Find Company Connections** ``` Discover relationships using records.search_by_relationship: 1. Start with target company ID 2. Use relationship_type="company_to_people" to find all contacts 3. Use records.search_by_relationship with "people_to_company" for reverse lookup 4. Show: person name, role, contact info, relationship strength Limit to 15 connections per company ``` #### **Account Mapping** ``` Build complete account map: 1. Use records.get_details for the target company 2. Use records.search_by_relationship to find all connected people 3. Use records.search_by_content to find related notes and activities 4. Use get-record-list-memberships to see which lists/pipelines they're in Present as organizational chart with contact details ``` ### 📈 Analytics & Reporting #### **Revenue Forecast (Optimized for Issue #469)** ``` Generate revenue forecast in batches: 1. Use records.search_by_timeframe for current quarter deals 2. Filter by stages: "Proposal", "Negotiation", "Closing" 3. Process in batches of 20 deals to avoid JSON truncation 4. Calculate totals by month and probability 5. Show top 10 deals by value, then ask for next batch ``` #### **Performance Dashboard** ``` Create performance summary using: 1. records.search_by_timeframe for deals closed this month 2. records.search_batch for multiple metrics in one query 3. Use records.get_attributes to identify available performance fields 4. Limit each metric to key numbers only Format: Won deals: X, Revenue: $Y, Pipeline: $Z, Activities: N ``` ### 🧹 Data Management #### **Find Incomplete Records** ``` Identify data gaps using records.search_advanced: 1. Resource type: companies 2. Filter where industry IS_EMPTY OR website IS_EMPTY 3. Employee count less than 1 OR greater than 1000000 (outliers) 4. Show: name, missing fields, last updated date Prioritize by deal value or engagement score ``` #### **Duplicate Detection** ``` Find potential duplicates: 1. Use records.search to get all companies 2. Use records.search_by_content for similar names/domains 3. Group by website domain for exact matches 4. Flag similar company names for manual review Process in batches of 25 companies to manage response size ``` ### 🎯 Batch Operations #### **Bulk Data Import** ``` Import multiple records using records.batch: 1. Operation type: "create" 2. Resource type: "companies" 3. Include all required fields: name, industry, website 4. Process in batches of 10 records for reliability 5. Use records.search_batch to verify successful creation Check each batch completion before proceeding ``` #### **Mass Updates** ``` Update multiple records efficiently: 1. Use records.search to find target records 2. Use records.batch with operation_type="update" 3. Specify resource_type and update fields 4. Limit to 15 records per batch for optimal performance 5. Verify updates with records.get_details for sample records ``` --- ## 🚨 Known Issues & Workarounds ### Issue #472: Tasks Showing "Unnamed" **❌ Avoid These Patterns:** ``` "Show me all my tasks" "List all tasks in the system" "Get all task data" ``` **✅ Use These Instead:** ``` "Use records.search_by_timeframe to find tasks created in last 7 days with content and title fields" "Get task details for overdue tasks, showing complete information" "Search for tasks linked to [Company Name] with full task data" ``` **When Creating Tasks:** ``` Always include both fields: - Content: "Detailed task description with context" - Title: "Short task name for lists" ``` ### Issue #469: JSON Response Truncation **❌ Avoid Large Requests:** ``` "Export all companies with complete details" "Show me everything about all deals" "Get comprehensive data export" ``` **✅ Use Pagination:** ``` "Show me companies 1-20 with key fields only" "Get the first 15 deals, then I'll ask for more" "Limit to essential fields: name, industry, value" ``` **Optimize Field Selection:** ``` Instead of: "Show complete company profiles" Use: "Show company name, industry, employee count, website only" ``` --- ## 🛠️ Troubleshooting Universal Tools ### When Tools Don't Work **1. Check Resource Type** ``` ❌ "Find some companies" ✅ "Use records.search with resource_type='companies' to find technology firms" ``` **2. Verify Available Fields** ``` Before complex queries, first use: "Use records.get_attributes with resource_type='companies' to see available fields" ``` **3. Build Incrementally** ``` Step 1: "Use records.search to find basic company data" Step 2: "From those results, filter for companies over 100 employees" Step 3: "Add industry filter for technology companies" ``` ### Performance Optimization **Limit Results:** ``` Add to every query: "Limit to 15 results" or "Show first 20 records" ``` **Use Specific Filters:** ``` ❌ "Find companies" ✅ "Find companies where industry='Technology' and employee_count>50" ``` **Batch Processing:** ``` "Process companies in groups of 10, starting with the first 10" ``` --- ## 📚 Quick Syntax Reference ### Universal Tool Parameters ```json { "resource_type": "companies|people|tasks|deals", "filters": { "filters": [ { "attribute": { "slug": "field_name" }, "condition": "equals|contains|greater_than|less_than|is_empty|is_not_empty", "value": "search_value" } ] }, "limit": 20 } ``` ### Common Conditions | **Condition** | **Use Case** | **Example** | | -------------- | --------------- | ---------------------------- | | `equals` | Exact matches | Industry equals "Technology" | | `contains` | Partial text | Name contains "Corp" | | `greater_than` | Numeric/date | Employee count > 50 | | `less_than` | Numeric/date | Deal value < 10000 | | `is_empty` | Missing data | Website is empty | | `is_not_empty` | Has data | Email address exists | | `starts_with` | Name prefixes | Company name starts with "A" | | `ends_with` | Domain suffixes | Email ends with ".com" | ### Resource Type Options - **companies**: Organizations and businesses - **people**: Contacts and individuals - **tasks**: To-dos and follow-up actions - **deals**: Opportunities and sales pipeline items - **lists**: Custom groupings and pipelines --- ## 🎯 Success Validation **✅ Your prompts are working correctly when:** - Results appear within 5 seconds - No "Unnamed" items in task lists - Response data isn't truncated with "..." - All requested fields are populated - Filtering returns expected record counts **🏆 You're a universal tools expert when you can:** - Build complex queries using multiple universal tools - Troubleshoot prompt issues by adjusting resource_type or conditions - Optimize large dataset queries with pagination - Create reliable batch operations for bulk data management --- **🔗 Related Documentation:** - [Getting Started Guide](getting-started.md) - Initial setup and first queries - [Troubleshooting Guide](troubleshooting.md) - Detailed problem solving - [Sales Playbook](playbooks/sales-playbook.md) - Ready-to-use sales prompts - [Operations Playbook](playbooks/operations-playbook.md) - Data management automation _All prompts in this guide are tested and validated with the current universal tools architecture (v2.0+)._