Agents Feedback

[BekahHW]

🚀 We just launched Continue Agents, our standardized AI workflows that combine prompts, rules, and tools for specific tasks!

We'd love your feedback on the new Agents feature. What workflows would you automate? Any feature requests or use cases we should consider.

• Agent Docs

[Mobious]

[BekahHW]

Can you share a little more details? Where are you running Continue agents? Can you link to the agent you're using? If you're running with the CLI, can you run it in TUI and share the output from your terminal?

[Mobious]

Maybe I'm misunderstanding what "agents" are? I'm just using the normal chat panel in "Agent" mode.

I'm running gpt-oss-120b on a custom server that emulates Ollama.

[Mobious]

Here's an example output.

The model calls ls successfully since tools is provided on the first call, but after the model gets the tool response, there is no tools list provided, so it ends up trying to call a hallucinated function that doesn't exist.

[BekahHW]

Ah yes, that's Agent Mode and not Agents, our standardized AI workflows that combine prompts, rules, and tools for specific tasks. I'm not able to move this conversation, so would you mind starting a new discussion with this information so I can delete it here?

[Mobious]

Got it. I Created a separate discussion.