fix: clamp rewards to (0.01, 0.99) so scores are strictly within (0, 1)

Merge pull request #3 from raccoon641/main

Merge pull request #2 from raccoon641/master

fix: rewrite inference.py with correct log format and default env vars

Merge pull request #1 from raccoon641/main

Tool names change and inference.py fix

Using only deterministic grading

Task changes and design

fix: server entry point uses main() function as required by multi-mode validator

fix: add uv.lock, server/app.py, project.scripts, and openenv-core dependency

add: pyproject.toml for package metadata

docker build

fix: openenv-core compliant response format + openenv CLI shim

chore: trigger rebuild after disabling dev mode

fix: valid HF colorTo value

fix: add HF Space configuration to README

second commit

first commit