# crmarena / crmarena_0053_knowledge_qa - taskset: [crmarena](https://harnessreport.com/tasks/crmarena.md) - difficulty: medium - category: crm - language: - runnable from the site: no - agent timeout: 600s ## Results by harness _none yet_ ## Instruction ``` # CRM Assistant Task You are an expert Salesforce CRM assistant helping an internal user (i.e., an employee of the same company). Your job is to answer the question below by reasoning step by step and querying a live Salesforce org using the `simple_salesforce` Python library. ## Approach Think step by step: 1. Identify which Salesforce objects and fields are relevant to the question 2. Write and execute SOQL/SOSL queries to retrieve the data you need 3. Inspect the results, refine your queries if needed 4. Write the final answer to `/app/answer.txt` ## Salesforce Access Connect using the credentials provided as environment variables: ```python import os from simple_salesforce import Salesforce sf = Salesforce( username=os.environ["SF_USERNAME"], password=os.environ["SF_PASSWORD"], security_token=os.environ.get("SF_SECURITY_TOKEN", ""), domain=os.environ.get("SF_DOMAIN", "login"), ) ``` Use SOQL for structured queries and SOSL for full-text search: ```python # SOQL — structured query result = sf.query_all("SELECT Id, Subject FROM Case WHERE Status = 'Open'") records = result["records"] # SOSL — full-text search across objects result = sf.search("FIND {Smith} IN NAME FIELDS RETURNING Contact(Id, FirstName, LastName)") records = result["searchRecords"] ``` ## Available Salesforce Objects The org contains these 16 objects (key fields listed): | Object | Key fields | |---|---| | `User` | Id, Alias, FirstName, LastName, Email, Phone | | `Account` | Id, FirstName, LastName, PersonEmail, Phone, ShippingCity, ShippingState, RecordTypeId | | `Contact` | Id, FirstName, LastName, Email, AccountId, OwnerId | | `Case` | Id, Subject, Description, Status, Priority, ContactId, AccountId, CreatedDate, ClosedDate, OwnerId, IssueId__c, OrderItemId__c | | `Knowledge__kav` | Id, Title, Summary, FAQ_Answer__c, UrlName | | `ProductCategory` | Id, Name, CatalogId | | `Product2` | Id, Name, Description, IsActive, External_ID__c | | `ProductCategoryProduct` | Id, ProductCategoryId, ProductId | | `Pricebook2` | Id, Name, Description, IsActive, ValidFrom, ValidTo | | `PricebookEntry` | Id, Pricebook2Id, Product2Id, UnitPrice | | `Order` | Id, AccountId, Pricebook2Id, Status, EffectiveDate | | `OrderItem` | Id, OrderId, Product2Id, PricebookEntryId, Quantity, UnitPrice | | `EmailMessage` | Id, Subject, TextBody, FromAddress, MessageDate, ParentId, ToIds | | `LiveChatTranscript` | Id, Body, AccountId, ContactId, CaseId, OwnerId, EndTime, LiveChatVisitorId | | `Issue__c` | Id, Name | | `CaseHistory__c` | Id, CaseId__c, Field__c (one of: 'Case Creation', 'Owner Assignment', 'Case Closed'), NewValue__c, OldValue__c, CreatedDate | ## Task-Specific Instructions - Use the information retrieved from the knowledge articles to answer the question in a concise manner ## Additional Context # Domain Details ## Quarters of the Year - Q1: January 1 to March 31 (both inclusive). - Q2: April 1 to June 30 (both inclusive). - Q3: July 1 to September 30 (both inclusive). - Q4: October 1 to December 31 (both inclusive). ## Seasons - Winter: December 1 to February 28/29 (both inclusive). - Spring: March 1 to May 31 (both inclusive). - Summer: June 1 to August 31 (both inclusive). - Autumn/Fall: September 1 to November 30 (both inclusive). ## Time Periods - Past 2 quarters: This refers to any timeframe spanning two quarters back from a specified `end_date`. That translates to a six-month period retrospectively from the `end_date`. - Issue Significantly More Than Other Months: This means there is a month where the number of cases reported are larger than all other months. ## Question How can grip enhancers benefit basketball players? ## Output Requirements Write your final answer to `/app/answer.txt`: - Single value: write it directly (e.g., `Agent A`) - Multiple values: comma-separated (e.g., `Agent A, Agent B, Agent C`) - No results: write `None` - Answer only — no explanations, no sentences, no punctuation beyond commas ```python from pathlib import Path Path("/app/answer.txt").write_text("your answer here") ``` ``` --- Harness Report runs agent harnesses from their GitHub repos on Harbor tasks and records every model call. Every page is also `.md` and `.json`; index: https://harnessreport.com/llms.txt · MCP: https://harnessreport.com/mcp