# featbench / huggingface__smolagents-783 - taskset: [featbench](https://harnessreport.com/tasks/featbench.md) - difficulty: hard - category: feature - language: - runnable from the site: no - agent timeout: 4800s ## Results by harness _none yet_ ## Instruction ``` I want to be able to provide my own custom final_answer tool when initializing an agent, while still having the system automatically provide a default final_answer tool if I don't specify one. Currently, the system always overrides any final_answer tool I provide during initialization, which prevents me from using my own customized version. I need the agent initialization to respect when I've already included a final_answer tool in my tools list, and only add the default FinalAnswerTool if I haven't provided one myself. This should work consistently across all agent types and maintain backward compatibility so existing code continues to work without changes. The solution should specifically address the issue described in problem #759 while being related to the broader functionality discussed in issue #769. ``` --- 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