Is agent-reward-bench safe to use?

agent-reward-bench has no known vulnerabilities on the latest version (0.1.2). DepScope rates its health at 0/100 (unknown risk).

What is the latest version of agent-reward-bench?

The latest version of agent-reward-bench is 0.1.2, released 2025-07-11.

Does agent-reward-bench have known vulnerabilities?

No known vulnerabilities on the latest version of agent-reward-bench.

Is agent-reward-bench deprecated?

No, agent-reward-bench is actively maintained. Latest release: 0.1.2 (2025-07-11).

What are alternatives to agent-reward-bench?

Popular alternatives to agent-reward-bench in PyPI include: boto3, packaging, urllib3, idna, requests.

What license does agent-reward-bench use?

License information for agent-reward-bench is not declared in the registry metadata.

agent-reward-bench

pypiv0.1.2

Official library for AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories

3 versions1 maintainers13 deps4 weekly dl

McGill-NLP/agent-reward-bench

/ 100

Health

insufficient data

agent-reward-bench has limited data (2/6 signals) — verify manually before use

Limited information available for this package
Moderate health score (40/100) — verify manually

Health breakdown0 – 100

10/25

maintenance

0/20

popularity

25/25

security

3/15

maturity

2/15

community

Vulnerabilities

none known

Health History

Dependency Tree

License Audit

Dependencies (13)

browsergym numpy pandas Pillow tqdm black;wheel;huggingface-hub;orjson;black;wheel;huggingface-hub;orjson;

API access

Get this data programmatically — free, no authentication.

curl https://depscope.dev/api/check/pypi/agent-reward-bench