Mahal Systems AgentVerse
Fits teams prototyping multi-agent orchestration rather than teams looking for a test set; the repo is a framework with a HumanEval runner, useful as a reference for recruit-plan-act-review loops.
What it asks the agent to do
- a recruited group of agents solves a coding problem together
- agents brainstorm and critique each other's answers
- agents collaborate on a tool-use puzzle such as 24-point
What to borrow from it
sandbox design
Caveats
It is a framework, not a benchmark; the tasks it ships (HumanEval and small puzzles) are saturated, and the last dated update is March 2024.
Details
| Workloads | multi agent, coding, general assistant |
|---|---|
| Interaction | autonomous long horizon |
| Environment | static |
| End state checked | No |
| Tools | code execution, optional external tools |
| Grading | unit tests |
| Size | Not published |
| Data | real public |
| Personal data | none |
| License | Apache-2.0 |
| Harness | own repo |
| Status | frozen |