Which integrations are available for this server?

Supports connecting to MySQL databases via an optional driver for profile storage and database instance profiling, including dialect/version metadata, metrics, and cross-instance relationships. Uses SQLite as the shared local archive store for instance profiles, observations, baselines, findings, audits, and metrics, allowing other MCPs to write to the same SQLite database through sqltk-store.

How do I use mcp-profile-hub?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-profile-hub show me the full profile for the orders database" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

mcp-profile-hub

档案中枢 · MCP Server

所有 MCP 的唯一数据落点 —— 也是「每次交付都在积累」这件事的物理载体。

安装（按需，不必装其他 MCP）

pip install mcp-profile-hub                        # 从 PyPI
pip install "./mcp-profile-hub[mysql]"             # 从源码（在工具链根目录执行）；要连库就补驱动，或 [postgres]

会自动带上它需要的共享库：sqltk-core（指纹/协议）、sqltk-store（真正干活的存储库）、sqltk-know（指标字典）。不会带其他 MCP。

Related MCP server: mcp-knowledge-server

挂到 MCP 客户端

{
  "mcpServers": {
    "profile-hub": {
      "command": "python",
      "args": ["-m", "profile_hub"],
      "env": { "DBT_STORE": "~/.dbtoolchain" }
    }
  }
}

按你的客户端（Claude Desktop / Cursor / Cline 等）说明，把这段合并到对应的 mcpServers 配置里。

命令行调试（不挂客户端也能用）：

mcp-profile-hub tools
mcp-profile-hub call <tool> --args '{"sql": "SELECT 1"}'

工具（24）

工具	作用
`compute_baselines`	从已有观测计算标量基线（每项指标一个常态值）。单点不算基线，样本不足跳过。
`compute_metric_baselines`	从 raw 指标样本计算分位数基线（p50/p95/p99）。样本不足的指标跳过，绝不编造。
`export_profile`	★ 导出脱敏知识包。
`finding_stats`	技术债总览（月报素材）：开放/已修复/误报率 + 明细清单。
`get_baseline`	读取已存在的基线；metric 不传则返回该实例全部基线。
`get_instance_profile`	实例全貌 —— 给 LLM 和客户看的完整画像（元数据 + 档位 + 最近观测 + 基线 + 技术债）。
`get_related_instances`	查与该实例关联的其他实例（同一业务的异构库、主从、上下游）。
`import_profile`	★ 导入知识包 —— 冷启动的第二条解法。
`link_cross_instance`	异构实例关联（同一业务跨多个引擎）—— 这是通用工具做不到的事。
`list_incidents`	事故单清单（含诊断结论与时间线）。
`list_instances`	列出档案库里登记过的实例（id / 方言 / 版本 / 标签 / 档位）。
`list_metrics`	列出某引擎支持的逻辑指标，方便确认能采到什么。
`metric_series`	读取一维指标时间序列（趋势/容量/异常检测的统一输入）。
`query_audit`	按时间倒序查审计留痕；instance_id 不传则跨实例返回。
`query_observations`	按时间倒序查观测记录；kind 不传则返回全部类型。
`record_audit`	写一条审计留痕：谁在什么时候对哪个实例做了什么。
`record_finding`	登记/累加一条体检发现（技术债账本）。重复出现自动 occurrences+1。
`record_observation`	写入一次观测。这是 Tier0 → Tier2 升级的唯一写入点
`set_baseline`	手工设定一个基线值（人工校准用）。
`store_stats`	档案库自身的情况：文件大小、各表行数。
`tier_of`	查询实例当前的能力档位。要求 Tier2 的能力会用它来明确报错，而不是悄悄降级
`top_sql`	出现次数最多的 SQL（一条 SQL 的一生）。
`transition_finding`	推进发现的生命周期：open→confirmed→fixing→fixed→verified / dismissed（状态机校验）。
`upsert_instance`	注册/更新实例元数据。url 会参与脱敏处理，密码绝不落盘。

闭环

upsert_instance 建档 → record_observation → compute_metric_baselines 分位数基线 → record_finding / transition_finding 技术债账本 → finding_stats 月报口径 → export_profile 知识包

Tier（冷启动保证）

tier_of 直接给出实例当前档位（tier1 = 一次观测 / tier2 = ≥2 次且跨 ≥7 天）。

边界：这个仓不做什么

本 MCP 不含任何判断逻辑，只负责把其他 MCP 产生的档案读写、聚合、导出。其他 MCP 不通过 MCP 协议调它，而是直接依赖 sqltk-store 这个库写同一个 SQLite —— 少一跳，且装不装 hub 都不影响记录。

开发

python tests/test_profile_hub.py      # 不需要 pytest，也不需要数据库

许可

This server cannot be deployed

Context All Your Agents Can ShareAgents share what they learn, hand off work, and hold it when you switch models. Start with 5 GB free.fulcradynamics.comAd

Maintenance

ActivityMaintained

ResponsivenessNo issues

Resources

Related MCP Connectors

Algenta MCP Server A
io.github.thyn-ai
Governed data discovery, exact queries, decisions, simulations, and runtime utilities over MCP.
Data Platforms Search RAG Systems
21
https://api.algenta.ai/mcp
n3tz AtlasOAuth
ai.n3tz
Governed company knowledge over MCP. Access by contract for authorized tenants.
Knowledge & Memory Security Search
1
https://atlas.n3tz.ai/mcp
dvtOAuth
dev.dvt
Dashboards as data. Author, validate, render, and share dvt dashboards from any MCP client.
Data Visualization
1
https://mcp.dvt.dev/mcp
OntoRamp Knowledge Cartographer
com.ontoramp
Knowledge coverage map and health score. Ingest docs into a governed knowledge graph via MCP.
Knowledge & Memory Documentation Access
2
https://ontoramp-knowledge-cartographer.fly.dev/mcp

Related MCP Servers

semantic-runtime
Knowledge & Memory RAG Systems
Chloride233
A
license
Not graded
quality
C
maintenance
Enables MCP clients to serve and query semantic models, providing tools for entity descriptions, metric lookups, context resolution, and operation validation for AI agents.
Updated 3 months ago (2026-07-31 12:52 UTC)
MIT
mcp-knowledge-server
Knowledge & Memory Testing & QA Tools Developer Tools
berns722
F
license
Not graded
quality
C
maintenance
Exposes a version-controlled knowledge base as MCP tools so agents can read merged identity/state summaries and next recommended actions at runtime, plus run deterministic schema validation and cross-reference link checks. This lets agents fetch live state and verify integrity via self-describing calls instead of re-reading raw files.
Updated a month ago (2026-09-16 06:05 UTC)
-
Memory Bridge
Knowledge & Memory RAG Systems
yanglinfeng
A
license
A
quality
A
maintenance
Enables MCP clients to store, recall, correct, and forget knowledge with per-sentence source bindings and a replayable retrieval traceId, so every answer can be traced to a specific document passage. Adds document-level version chains that retire superseded content from default retrieval, tombstone-based forgetting, three-layer user/namespace/scope isolation with full operation auditing, and deterministic calculator/date tools.
Updated 10 days ago (2026-10-16 03:53 UTC)
10
Apache 2.0
benchmark-test-weakening
AI & Machine Learning Testing & QA Tools Developer Tools
pooriaarab
A
license
Not graded
quality
A
maintenance
Enables MCP clients to list and inspect coding-agent benchmark tasks, score recorded answers, run evaluations, and retrieve data provenance.
Updated 9 days ago (2026-10-16 17:42 UTC)
295 npm
MIT