Add AgentEnv to the dataset libraries

[earakely-scale]

Adds agentenv to DATASET_LIBRARIES_UI_ELEMENTS, next to the environment frameworks #2441 registered (Harbor, Verifiers, OpenEnv, NeMo Gym).

AgentEnv (scaleapi/agentenv-framework) is an open-source framework for building dynamic RL environments and running and scoring agents in them. AgentEnv environments and their recorded runs are published to the Hub as datasets tagged rl-environment and agentenv, for example earakely-scale/PortSimEnv-AgentEnv and earakely-scale/agentenv-hf-smoke.

Checked in packages/tasks: pnpm run check, lint:check, format:check and test (64 passed).

Companion PRs that register AgentEnv: #2638 (the library entry: badge and button), huggingface/dataset-viewer#3463 (the agentenv tag and its "Use this dataset" command), huggingface/hub-docs#2870 (the framework table).

🤖 Generated with Claude Code

---

Note

Low Risk
Adds a static UI metadata entry only; no runtime logic or security-sensitive paths change.

Overview
Registers AgentEnv as a Hub dataset library by adding an agentenv entry to DATASET_LIBRARIES_UI_ELEMENTS in dataset-libraries.ts, with display label, GitHub repo (scaleapi/agentenv-framework), and docs URL—alongside other RL environment frameworks (Harbor, Verifiers, OpenEnv, NeMo Gym).

The DatasetLibraryKey type picks up the new key automatically. Hub UI for badges and “Use this dataset” flows are handled in companion PRs elsewhere in the Hugging Face stack.

^{Reviewed by Cursor Bugbot for commit 8eddb31. Bugbot is set up for automated code reviews on this repo. Configure here.}