Skip to main content
  • Context window: 128,000 tokens
  • Model type: Dense
  • Supported modes: Agentic, Chat

​
Summary

Malibu Agent is a new family of models from Poolside that is capable of autonomously solving complex coding tasks over multiple rounds of tool use without human intervention. It calls tools to gather context, perform edits, and run applications in order to solve real world coding problems within large codebases. Malibu Agent is designed to be run within the Poolside Agent Framework, accessed within Poolside’s editor extension, CLI and web interface.

​
Improvements

Contrary to Poolside’s chat model, Malibu 2.1, Malibu Agent is trained to perform multiple rounds of tool calls in order to solve issues or respond to questions. You should give it a direct problem to solve, with context such as stack traces or file references; it will then read your codebase to understand the issue, perform edits, and verify that its changes solve the issue. Some example prompts might look like:
  • Please figure out why step_events is null in @trajectory.jsx.
  • Can you debug this 500 request? <insert your stacktrace>.
  • Can you give me an overview of how ResponseProvider works and how it interacts with TrajectorySource? Where does it store trajectories?
Given a request, you’ll probably see Malibu Agent perform the following operations:
  • Read files and folders to understand project structure.
  • Perform edits to files.
  • Run programs (such as linters or compilers) to validate its edits, or even run tests and standalone web servers to validate its fixes.

​
Tips for prompting

  • Give Malibu Agent as much context as possible if you have a specific bug, including stacktraces and references to files that may be useful.
  • If Malibu Agent is unable to solve an issue, break your request down into smaller, incremental changes, or be more direct (“Make X change” instead of “implement this feature”).

​
Known issues

  • Malibu Agent can be over-reactive to simple commands like “hi” or “how are you today?” - as it’s trained to solve issues, it will interpret these simple commands as problems to solve and attempt to solve them. This is a known issue and will be improved in future releases.
  • Malibu Agent can make edits to files unrelated to the task at hand; if this occurs, use the checkpointing feature to roll back its changes.
  • Malibu Agent can sometimes create test scripts to validate its changes, but forget to delete them. Check to make sure you don’t accidentally commit these to version control.