{"$schema":"https://wellknown.network/schemas/agent-record-v1.json","schemaVersion":"1","id":"ag_n38tgbhv4sjk","handle":"aika-research-mcp","url":"https://wellknown.network/agents/aika-research-mcp","links":{"self":"https://wellknown.network/agents/aika-research-mcp/record.json","html":"https://wellknown.network/agents/aika-research-mcp","markdown":"https://wellknown.network/agents/aika-research-mcp/record.md","api":"https://wellknown.network/api/v1/agents/aika-research-mcp","status":"https://wellknown.network/api/v1/agents/aika-research-mcp/status","claim":"https://wellknown.network/agents/aika-research-mcp/claim","claimApi":"https://wellknown.network/api/v1/claims","claimDescriptor":"https://wellknown.network/agents/aika-research-mcp/claim.json","badge":"https://wellknown.network/agents/aika-research-mcp/badge.svg","openapi":"https://wellknown.network/openapi.json","history":"https://wellknown.network/api/v1/agents/aika-research-mcp/history","tools":"https://wellknown.network/api/v1/agents/aika-research-mcp/tools"},"ard":{"identifier":"urn:air::server:aika-research-mcp","type":"application/mcp-server-card+json"},"kind":"mcp_server","declared":{"name":"aika-research-mcp","summary":"Evidence-driven AI compute industry-chain research MCP server and local SQLite CLI.","description":"# AIKA\n\n基于 AI 算力产业链的投研 Agent 问答系统 —— 可嵌入第三方 Agent 的 MCP/Skill 投研插件\n\n项目简介：面向 AI 算力产业链投研场景，构建可追溯、有证据引用的 Agent 问答系统，并以 MCP Server + Skill 插件形态嵌入 Claude Code 等第三方 Agent 工作流。覆盖 AI 服务器、芯片、光模块、液冷、数据中心等核心主题，支持产业链分析、公司对比、风险识别、证据引用和研究报告生成。\n\n技术栈：Python + LangGraph + RAG + 知识图谱 + MCP\n\n项目亮点：\n• 投研知识库构建：整合 61 份 PDF 文档，构建 RAG 知识库、Claim / Evidence 证据体系、Segment Dossier 和实体关系图谱，形成结构化投研知识底座\n• 混合问答链路设计：设计\"Agent + 知识图谱 + RAG\"混合问答流程，实现任务规划、多路检索、证据重排、答案生成与支撑性校验\n• MCP/Skill 插件化：将投研能力封装为 MCP Server 与 Skill，支持一键安装至 Claude Code 等宿主 Agent，无需 Docker 或 Web 即可在 Agent 对话中直接调用投研工具\n• RAG 检索评测与系统验证：基于 50 条测试用例完成 BM25、语义搜索及 RRF 融合检索评测，并通过 LLM 自动标注增强降低 unjudged 率；同步完成 15 个投研问题 QA smoke test 与 15 个 Agent 任务样例测试，验证端到端可用性\n\n---\n\nAIKA 是一个面向 AI 算力产业链的证据驱动智能问答与投研 Agent 系统。公开版主路径是本地 SQLite + MCP/Skill：用户不需要启动 Docker、PostgreSQL、Neo4j 或 Web，就可以在宿主 Agent 中调用投研工具。FastAPI + React Web 版保留为本地开发测试台，用于调试对话、证据卡、图谱、反馈和 Agent 任务视图，不作为公网多用户产品入口。\n\n### Quick Start\n\n用户可以直接从 PyPI 安装并运行本地 sample 流程：\n\n```bash\npip install aika-research-mcp\naika init --sample\naika build-index\naika doctor\naika demo\n```\n\n源码开发环境可以使用 `uv run` 执行同一套公开版 CLI：\n\n```bash\nUV_CACHE_DIR=/tmp/uv-cache uv run aika init --sample\nUV_CACHE_DIR=/tmp/uv-cache uv run aika build-index\nUV_CACHE_DIR=/tmp/uv-cache uv run aika doctor\nUV_CACHE_DIR=/tmp/uv-cache uv run aika demo\n```\n\n### Local MCP Quick Start\n\n公开版 AIKA 可以作为本地 MCP Server + Skill 接入 Claude Code 或 Codex。用户不需要手写 `.mcp.json`、`~/.claude.json`、`~/.codex/config.toml`，也不需要手动复制 `skills/aika-research/SKILL.md`。\n\nClaude Code 用户级安装：\n\n```bash\naika mcp install --host claude-code --scope user --with-skill\nclaude\n/mcp\n```\n\nCodex CLI 和 Codex IDE extension 共享 MCP 配置，用户级安装一次即可复用：\n\n```bash\naika mcp install --host codex --scope user --with-skill\ncodex\n/mcp\n```\n\n也可以把配置和 Skill 写入当前项目。Codex 会在该项目被信任后加载项目级 `.codex/config.toml` 和 `.codex/skills/aika-research/`：\n\n```bash\naika mcp install --host codex --scope project --with-skill\n```\n\n安装完成后，可以在宿主 Agent 中显式使用 AIKA Skill：\n\n```text\n使用 $aika-research 分析液冷产业链，要求保留 citation id。\n```\n\n### Skill…","publisher":{"name":"AIKA Contributors","url":null},"homepage":"https://github.com/To-sam-akf/AIChainQASysterm","repository":"https://github.com/To-sam-akf/AIChainQASysterm","version":"0.1.1","license":"Copyright (c) AIKA Contributors. All rights reserved.","protocols":["mcp"],"tags":["ai","compute","aika","graphrag","mcp","sqlite","research"],"pricing":null,"endpoints":[{"url":"pypi:aika-research-mcp","type":"package_pypi","auth":null,"probeable":false}],"skills":null,"tools":null,"extra":null,"attribution":{"kind":"pypi","name":"pypi","license":"pypi","repoUrl":"pypi","summary":"pypi","version":"pypi","description":"pypi","homepageUrl":"pypi","publisherName":"pypi"}},"derived":{"capabilities":[{"slug":"data.database","name":"Databases","confidence":1,"provenance":"declared"}],"categories":["data"],"language":"en"},"observed":{"status":"unknown","statusReason":"Distributed as a package to run locally; no network endpoint to check.","lastOkAt":null,"lastProbedAt":null,"statusComputedAt":null,"reliability30d":null,"latestObservations":[],"tools":null,"package":{"name":"aika-research-mcp","registry":"pypi","observedAt":"2026-11-08T22:25:09.875Z","publishedAt":"2026-06-19T08:30:09.878809Z","latestVersion":"0.1.1"},"toolSurface":null,"endpointFacts":[]},"verification":{"claimed":false,"claimedAt":null,"proofs":[]},"provenance":{"sources":[{"source":"pypi","key":"aika-research-mcp","url":"https://pypi.org/project/aika-research-mcp/","firstSeenAt":"2026-09-09T08:21:42.948Z","fetchedAt":"2026-11-08T22:24:13.845Z","normalizedAt":"2026-11-08T22:24:13.845Z"}]},"firstSeenAt":"2026-09-09T08:21:42.948Z","updatedAt":"2026-11-08T22:25:09.875Z"}