# AttributeError: 'str' object has no attribute 'get' after the task is finished

## TL;DR

[sonichi (commenter)]: Same root cause as I said above. The local LLM does not allow empty content, so when the model returns an empty message the agent crashes with AttributeError: 'str' object has no attribute 'get'. The workaround is to set default_auto_reply to a non-empty string on your agents, so there is always content to work with when the model sends nothing back.

Use this when you run into the situation in the title.

**When not to use this skill:** unrelated tasks. It covers only the procedure above.

## Compatibility

The steps above apply to the commands named in them. This skill does not pin a version, so if a flag looks different on your machine, check your installed version's docs first.

## Details

[https://github.com/microsoft/autogen/issues/279#issuecomment-1773528968](url)

Context: GitHub issue microsoft/autogen#279 (closed, 44 comments): Hi! First, thank you to all the devs who made autogen, it is so cool! Unfortunately I have a persistent error (AttributeError: 'str' object has no attribute 'get') It pops up sometimes before the Assistant is finished (when it writes python code for example), but when the task is primitive ( "come up with 5 possible names for SMMA that helps companies get leads") it manages to get it done and then spits out this error. `user_proxy (to assistant): come up with 5 possible names for SMMA that helps companies get leads -------------------------------------------------------------------------------
