Mahal Systems E-Commerce Bench
Fits agents that operate a business over a long horizon. Borrow the deterministic demand and negotiation kernels (the LLM only renders dialogue) and the multi-axis scoring with assets as the headline.
What it asks the agent to do
- run online stores for a simulated year (pricing, stocking, fulfilment, returns)
- negotiate with suppliers and avoid fraudulent ones
- manage cash flow under escrow delays and demand shocks
What to borrow from it
sandbox designgrading design
Caveats
One scenario with five episodes per model; very long runs; outcome is a single number that hides which step failed.
Details
| Workloads | retail, sales, planning, finance |
|---|---|
| Interaction | autonomous long horizon |
| Environment | api sandbox |
| End state checked | Yes |
| Tools | 18 merchant tools over 6, 886 products and 576 suppliers |
| Grading | programmatic check |
| Size | 1 |
| Data | mixed |
| Personal data | none |
| License | Apache-2.0 |
| Harness | own repo |
| Status | active |