Spaces:
Runtime error
Runtime error
| """Red & Blue agent layer for OpenRange. | |
| Exports: | |
| - RangeAgent: Protocol for any compatible agent | |
| - EpisodeResult, EpisodeMetrics: Trajectory/metrics dataclasses | |
| - LLMRangeAgent: LiteLLM-powered agent (any model) | |
| - ScriptedAgent, ScriptedRedAgent, ScriptedBlueAgent: Fixed-sequence replay agents | |
| - HumanAgent: Interactive stdin/stdout agent | |
| - run_episode: Orchestration loop | |
| - evaluate: Multi-episode evaluation harness | |
| - extract_command: LLM output -> shell command parser | |
| """ | |
| from open_range.agents.protocol import EpisodeMetrics, EpisodeResult, RangeAgent | |
| from open_range.agents.parsing import extract_command | |
| from open_range.agents.replay_agent import ( | |
| ScriptedAgent, | |
| ScriptedBlueAgent, | |
| ScriptedRedAgent, | |
| ) | |
| from open_range.agents.human_agent import HumanAgent | |
| from open_range.agents.llm_agent import LLMRangeAgent | |
| from open_range.agents.episode import run_episode | |
| from open_range.agents.eval import evaluate | |
| from open_range.agents.prompts import BLUE_SYSTEM_PROMPT, RED_SYSTEM_PROMPT | |
| __all__ = [ | |
| "RangeAgent", | |
| "EpisodeResult", | |
| "EpisodeMetrics", | |
| "LLMRangeAgent", | |
| "ScriptedAgent", | |
| "ScriptedRedAgent", | |
| "ScriptedBlueAgent", | |
| "HumanAgent", | |
| "run_episode", | |
| "evaluate", | |
| "extract_command", | |
| "RED_SYSTEM_PROMPT", | |
| "BLUE_SYSTEM_PROMPT", | |
| ] | |