{"id":23932703,"url":"https://github.com/aws-samples/multiagent-collab-scenario-benchmark","last_synced_at":"2025-09-11T15:32:07.912Z","repository":{"id":267490544,"uuid":"884450047","full_name":"aws-samples/multiagent-collab-scenario-benchmark","owner":"aws-samples","description":"Benchmarking data and script used for LLM multi-agent collaboration systems from AWS Bedrock Agents Science team.","archived":false,"fork":false,"pushed_at":"2024-12-10T15:59:34.000Z","size":98,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":3,"default_branch":"main","last_synced_at":"2024-12-10T17:30:53.659Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit-0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/aws-samples.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":"CODE_OF_CONDUCT.md","threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-11-06T19:17:30.000Z","updated_at":"2024-12-10T15:59:37.000Z","dependencies_parsed_at":"2024-12-10T17:43:09.249Z","dependency_job_id":null,"html_url":"https://github.com/aws-samples/multiagent-collab-scenario-benchmark","commit_stats":null,"previous_names":["aws-samples/multiagent-collab-scenario-benchmark"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aws-samples%2Fmultiagent-collab-scenario-benchmark","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aws-samples%2Fmultiagent-collab-scenario-benchmark/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aws-samples%2Fmultiagent-collab-scenario-benchmark/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/aws-samples%2Fmultiagent-collab-scenario-benchmark/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/aws-samples","download_url":"https://codeload.github.com/aws-samples/multiagent-collab-scenario-benchmark/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":232657441,"owners_count":18556845,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-01-06T00:29:17.842Z","updated_at":"2025-01-06T00:29:25.047Z","avatar_url":"https://github.com/aws-samples.png","language":"Python","funding_links":[],"categories":["Building"],"sub_categories":["Benchmarks"],"readme":"## Multi-agent Collaboration Scenario Benchmarking\n\nThis repository contains benchmarking material from the AWS Bedrock Agents multi-agents collaboration technical report: \"Towards Effective GenAI Multi-Agent Collaboration: Design and Evaluation for Enterprise Applications\". The technical report is here on arXiv: https://arxiv.org/abs/2412.05449.\n\n### Data\n\nBenchmarking data is in the `datasets` directory where there are 30 hypothetical scenarios for three domains: travel planning, mortgage financing, and software development. \n\nEach entry in the scenarios file contains:\n- `scenario`: The user background and goals.\n- `input_problem`: A description of the problem to be solved by the agent.\n- `assertions`: A list of assertions that must be true to judge the interaction between user and the agent. \n\nIn each dataset, there is also a `agents.json` file that contains the agent's name and description, as well as their corresponding tools. The scenarios are collected based on these agent profiles and tool schemas.\n\n### Pre-requisites\n\nCreate a Python 3.12 virtual environment and install requirements in `requirements.txt`.\n\nNext, prepare the conversations that you want to benchmark. Each conversation should be in its own JSON file titled `conversation_0.json`, `conversation_1.json`, etc. where the index corresponds to the scenario index. The `conversation_{i}.json` file should be formatted as follows:\n\n```\n{\n    \"trajectories\": {\n        \"agent_id_1\": [\n            {\n                \"role\": null, # null, User, Action, Observation\n                \"source\": \"\", # agent_id of the agent who sent this message\n                \"destination\": \"\", # agent_id of the user who received this message\n                \"content\": \"\", # content of the message\n                \"actions\": [], # list of action objects executed by the agent\n                \"observation\": null, # observation of the agent\n            }\n        ],\n        \"agent_id_2\": [...],\n        ...\n    }\n}\n```\n\nSee `sample_conversations` for examples.\n\n\n### How to use \n\nFirst, export any environment variables needed for LLM providers (Bedrock, OpenAI, Anthropic, etc) to support the LLM judge. See [LiteLLM Providers](https://docs.litellm.ai/docs/providers) for setting up LLMs.\n\nRun the benchmarking script on a sample travel conversation:\n\n```\n{export env variables}\n\npython -m src.benchmark\n```\n\nCustomize the benchmarking parameters as needed:\n```\npython -m src.benchmark \\ \n    --dataset_dir \u003cpath_to_dataset\u003e  \\\n    --scenario_filename \u003cfilename of scenarios\u003e \\\n    --conversations_dir \u003cpath_to_conversations\u003e \\\n    --llm_judge_id \u003cLiteLLM llm_judge_id\u003e \\\n```\n\n\n\n## Security\n\nSee [CONTRIBUTING](CONTRIBUTING.md#security-issue-notifications) for more information.\n\n## License\n\nThis library is licensed under the MIT-0 License. See the LICENSE file.\n\nThe dataset is licensed under the CC-BY-4.0 license.\n\n## Citation\n\nIf you have found our work useful, please cite the technical report:\n\n```\n@misc{shu2024effectivegenaimultiagentcollaboration,\n      title={Towards Effective GenAI Multi-Agent Collaboration: Design and Evaluation for Enterprise Applications}, \n      author={Raphael Shu and Nilaksh Das and Michelle Yuan and Monica Sunkara and Yi Zhang},\n      year={2024},\n      eprint={2412.05449},\n      archivePrefix={arXiv},\n      primaryClass={cs.CL},\n      url={https://arxiv.org/abs/2412.05449}, \n}\n```\n## Core Contributors\n* [Raphael Shu](https://github.com/zomux)\n* [Nilaksh Das](https://github.com/nilakshdas)\n* [Michelle Yuan](https://github.com/forest-snow)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faws-samples%2Fmultiagent-collab-scenario-benchmark","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Faws-samples%2Fmultiagent-collab-scenario-benchmark","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Faws-samples%2Fmultiagent-collab-scenario-benchmark/lists"}