Mahal Systems USACO
A cost-tracked algorithmic baseline available on HAL; useful as a sanity check for reasoning-heavy coding agents, not for product engineering.
What it asks the agent to do
- solve a USA Computing Olympiad problem in Python against hidden tests
- use retrieval of similar problems or reflection to improve a solution
What to borrow from it
baseline
Caveats
Olympiad puzzles with public solutions online; no repository or tools; license is not stated in the repo.
Details
| Workloads | coding |
|---|---|
| Interaction | single turn |
| Environment | static |
| End state checked | No |
| Tools | none |
| Grading | unit tests |
| Size | 307 |
| Data | real public |
| Personal data | none |
| License | unknown |
| Harness | inspect_evals, hal, own repo |
| Status | active |