0

Search ENV RL Env (Prime Intellect)

Fresh

MCP Search Environment Example

Type
RL Env
Publisher
Prime Intellect
Runtime
multi-turn
License
unknown
Size
v0.1.0
Published
Dec 2025
Canonical
app.primeintellect.ai/dashboard/environments/primeintellect/mcp-search-env

Cite

Notes

Only stored in your browser.

Attribution

Attribution policy →

mcp-search-env

Source Code

Overview

  • Environment ID: mcp-search-env
  • Short description: Example environment using vf.MCPEnv for MCP server integration
  • Tags: MCP, Tools

This environment demonstrates how to use the first-class MCPEnv from verifiers.envs.experimental.

Datasets

  • Primary dataset(s): N/A
  • Source links: N/A
  • Split sizes: N/A

Task

  • Type: <multi-turn | tool use>
  • Rubric overview: N/A

Quickstart

Run an evaluation with default settings:

prime eval run mcp-search-env

Configure model and sampling:

prime eval run mcp-search-env   -m gpt-4.1-mini   -n 1 -r 1

Notes:

  • Use -a / --env-args to pass environment-specific configuration as a JSON object.

Environment Arguments

Document any supported environment arguments and their meaning. Example:

ArgTypeDefaultDescription
max_examplesint-1Limit on dataset size (use -1 for all)

Metrics

MetricMeaning
rewardMain scalar reward (weighted sum of criteria)
judge_rewardLLM judge score (1.0 if correct, 0.0 if not)