功能：
工作流是一个通过google-search和 githubapi 工具获取开源项目最近30天各项活动信息的总结；
输入：
pytorch/pytorch
输出：
agent追踪日志：
{
"model": {
"completion_params": {},
"mode": "chat",
"model": "Qwen/Qwen3-30B-A3B-Instruct-2507",
"model_type": "llm",
"provider": "langgenius/siliconflow/siliconflow",
"type": "model-selector"
},
"tools": [
{
"credential_id": "a4c514bd-ac5e-41c2-b212-393dc083b1e1",
"enabled": true,
"extra": {
"description": "A tool for performing Google SERP searches and extracting snippets and web pages. Input should be a search query."
},
"parameters": {
"country_code": null,
"language_code": null,
"query": null
},
"provider_name": "langgenius/google/google",
"provider_show_name": "langgenius/google/google",
"settings": {
"as_agent_tool": {
"type": "constant",
"value": false
}
},
"tool_description": "A tool for performing Google SERP searches and extracting snippets and web pages. Input should be a search query.",
"tool_label": "Google Search",
"tool_name": "google_search",
"type": "builtin"
},
{
"enabled": true,
"extra": {
"description": "Fetches issues from a GitHub repository, supporting status filtering, sorting, and pagination options"
},
"parameters": {
"direction": null,
"owner": null,
"per_page": null,
"repo": null,
"sort": null,
"state": null
},
"provider_name": "langgenius/github/github",
"provider_show_name": "langgenius/github/github",
"settings": {},
"tool_description": "Fetches issues from a GitHub repository, supporting status filtering, sorting, and pagination options",
"tool_label": "Repository Issues",
"tool_name": "github_repository_issues",
"type": "builtin"
},
{
"enabled": true,
"extra": {
"description": "Fetches Pull Requests from a GitHub repository, supporting status filtering, sorting, and pagination options"
},
"parameters": {
"direction": null,
"owner": null,
"per_page": null,
"repo": null,
"sort": null,
"state": null
},
"provider_name": "langgenius/github/github",
"provider_show_name": "langgenius/github/github",
"settings": {},
"tool_description": "Fetches Pull Requests from a GitHub repository, supporting status filtering, sorting, and pagination options",
"tool_label": "Repository Pull Requests",
"tool_name": "github_repository_pulls",
"type": "builtin"
},
{
"enabled": true,
"extra": {
"description": "Fetches releases from a GitHub repository, supporting pagination options"
},
"parameters": {
"owner": null,
"per_page": null,
"repo": null
},
"provider_name": "langgenius/github/github",
"provider_show_name": "langgenius/github/github",
"settings": {},
"tool_description": "Fetches releases from a GitHub repository, supporting pagination options",
"tool_label": "Repository Releases",
"tool_name": "github_repository_releases",
"type": "builtin"
}
],
"instruction": "You are a professional open-source community monitoring and analysis agent. Your task is to use available search tools (such as Google Search) and corresponding GitHub tools to query activity information for user-specified GitHub repositories within the last 30 days, and generate a structured, accurate, and readable summary report in Simplified Chinese based on the search results, providing jump links for each key content in the report.\nPlease execute according to the following requirements:\nActively collect the latest information related to the target repository from search results, with the time frame strictly limited to the last 30 days. The scope of information includes but is not limited to:\nGitHub Repository Internal Activities: Issues, Pull Requests, Releases, Tags, Milestones, Roadmap, Projects, Discussions.\nExternal Related Content: Technical blogs, official or community announcements, technical conferences, industry news, external technical articles, core member statements or blogs, security incidents or vulnerability disclosures.\nOrganize and analyze search results:\nDeduplicate, filter noise, and remove irrelevant content.\nAutomatically ignore information older than 30 days from the current date.\nSimplify complex content into clear, concise descriptions.\nAutomatically categorize content into corresponding categories.\nExtract core information such as latest progress, important features, new functionalities, critical fixes, important discussions, release announcements, and their corresponding links.\nWhen search results contain time information (e.g., '3 days ago', '2 weeks ago', 'Nov 2024'), prioritize determining if it falls within the last 30 days.\n\nOutput the final result in the following structured format, with links, maintaining clarity, conciseness, and readability:\nThis Period's GitHub Activity Summary (Repository: {Repository Name}, Time Frame: Last 30 Days)\nLatest Major Developments:\nImportant New Features and Links:\nCritical Bug Fixes and Links:\nImportant PRs and Issues and Links:\nPR #xxxx: ...\nIssue #xxxx: ...\nRelease / Tag Summary and Links (Within Last 30 Days):\nVersion: ...\nKey Changes: ...\nConferences / Blogs / News / External Dynamics and Links (Within Last 30 Days):\nImportant Conferences: ...\nBlogs and Technical Articles: ...\nIndustry or Community News: ...\nRoadmap / Future Trends Links:\nComprehensive Summary: Combine all the above key information to provide a brief overview of the repository's activities in the past 30 days, in 400-500 words.\nOutput Requirements: Use Simplified Chinese. The content must be true and accurate, clearly structured, not fabricated, and must be information from the last 30 days. If a certain type of information has no content within the last 30 days, please mark it as 'No relevant content available'.",
"query": "pytorch/pytorch",
"maximum_iterations": 20
}
输出日志：
{
"text": "{\n \"action\": \"github_repository_pulls\",\n \"action_input\": {\n \"owner\": \"pytorch\",\n \"repo\": \"pytorch\",\n \"state\": \"all\",\n \"sort\": \"updated\",\n \"direction\": \"desc\",\n \"per_page\": 10\n }\n}",
"usage": {
"completion_price": "0",
"completion_price_unit": "0",
"completion_tokens": 0,
"completion_unit_price": "0",
"currency": "RMB",
"latency": 0,
"prompt_price": "0",
"prompt_price_unit": "0",
"prompt_tokens": 0,
"prompt_unit_price": "0",
"time_to_first_token": null,
"time_to_generate": null,
"total_price": "0.0015141",
"total_tokens": 1965
},
"files": [],
"json": [
{
"data": {
"action_input": "",
"action_name": "",
"observation": "",
"thought": "{\n \"action\": \"github_repository_pulls\",\n \"action_input\": {\n \"owner\": \"pytorch\",\n \"repo\": \"pytorch\",\n \"state\": \"all\",\n \"sort\": \"updated\",\n \"direction\": \"desc\",\n \"per_page\": 10\n }\n}"
},
"error": null,
"id": "c06f0b6f-c92d-46e2-a850-175461b252f5",
"label": "ROUND 1",
"metadata": {
"currency": "RMB",
"elapsed_time": 5.772278695000068,
"finished_at": 1149.064862912,
"started_at": 1143.292584565,
"total_price": "0.0015141",
"total_tokens": 1965
},
"node_id": "17657657450640",
"parent_id": null,
"status": "success"
},
{
"data": {
"action": "{\n \"action\": \"github_repository_pulls\",\n \"action_input\": {\n \"owner\": \"pytorch\",\n \"repo\": \"pytorch\",\n \"state\": \"all\",\n \"sort\": \"updated\",\n \"direction\": \"desc\",\n \"per_page\": 10\n }\n}",
"thought": "{\n \"action\": \"github_repository_pulls\",\n \"action_input\": {\n \"owner\": \"pytorch\",\n \"repo\": \"pytorch\",\n \"state\": \"all\",\n \"sort\": \"updated\",\n \"direction\": \"desc\",\n \"per_page\": 10\n }\n}"
},
"error": null,
"id": "62b9c401-e835-4110-ba52-21965a04b983",
"label": "Qwen/Qwen3-30B-A3B-Instruct-2507 Thought",
"metadata": {
"currency": "RMB",
"elapsed_time": 5.771518316000083,
"finished_at": 1149.064601222,
"icon": "0179a4422f60cc28de38c7b3395b8504c9a7516c050d355849784a46bcea3b91.svg",
"icon_dark": null,
"provider": "langgenius/siliconflow/siliconflow",
"started_at": 1143.293083623,
"total_price": "0.0015141",
"total_tokens": 1965
},
"node_id": "17657657450640",
"parent_id": "c06f0b6f-c92d-46e2-a850-175461b252f5",
"status": "success"
},
{
"data": []
}
]
}