Skip to main content
Glama
GelortZCY

research-mentor-rag

by GelortZCY
OverviewSchemaRelated ServersScoreDiscussions
Python
Hybrid

Research Mentor RAG

Research Mentor RAG 不是一个等人提问的论文搜索框。它是一套装进 AI 工作流里的科研导师系统：1200 篇论文被拆成 39,238 条可调用的研究记录，其中 9,697 条是阶段化方法。模型做选题、提出假设、设计实验、制定统计方案、写论文、回应审稿或准备复现时，系统会先检索相似研究和失败边界，再允许它给出判断。

真正负责调度的是 Research Mentor Workflows Skill。它识别当前科研阶段，规定每一步要查哪些 RAG 检索头、调用哪些专业能力、产出什么制品，以及何时必须纠错或停在 HOLD。配套的 12-Skill 能力合同覆盖多源文献搜索、全文与图表读取、引文核验、假设生成、实验设计、实验日志、统计分析、论文写作、系统综述和审稿红队。总控 Skill 不会用一段泛泛建议代替这些专业流程，而是把任务路由给对应能力。

运行时通过 6 个 MCP 工具把这套能力直接交给模型：research_bootstrap 加载经验包，research_rag_query 定向查证据，research_route 执行阶段多头检索，research_mentor 组合 RAG 与专业 Skill，research_frontier_search 实时查询 OpenAlex 和 Crossref，research_controller 自动判断阶段并组装回答前的证据包。Codex Hook 会在检测到科研请求时提前触发总控。到了选题阶段，它要求同时查看本地最近邻和实时前沿；到了实验阶段，它检索相似实验的变量、对照、测量和停止规则；到了写作阶段，它拉取论证结构、图表证据和常见审稿风险。检索结果、采用或拒绝的节点、证据边界、纠错条件和下一步会一起注入模型上下文。

底层检索使用 SQLite FTS5，不需要嵌入模型。L0–L5 六层结构把原子记录、单篇论文、方法积木、跨论文经验、科研阶段和完整研究流程连在一起，既能快速给出全局路线，也能一路下钻到具体经验。当前公开版本是 FAST80 / PROVISIONAL：它已经能承担导航、方法选择和流程约束，但不会把尚未完成的正式双增量审计写成最终科学验证。

现在包含什么

  • 1200 篇冻结论文对应的经验记录；

  • 39,238 个原子记录与方法记录，其中 9,697 个是阶段化方法；

  • L0 原子证据 → L1 单篇论文 → L2 方法积木 → L3 跨论文经验 → L4 科研阶段 → L5 完整流程；

  • 7 个阶段：novelty、hypothesis、experiment、statistics、writing、review、reproducibility；

  • 6 个 MCP 工具：bootstrap、RAG query、route、mentor、frontier search、controller；

  • OpenAlex + Crossref 实时前沿发现，用于新颖性和研究缺口初筛；

  • Codex UserPromptSubmit Hook 示例，可在科研请求进入模型前注入阶段检索结果。

Related MCP server: Hybrid-Gemini MCP

快速开始

要求 Python 3.11+，其余运行依赖均来自标准库。

python scripts/smoke_test.py

直接检索实验设计经验：

python skill/research-mentor-workflows/scripts/fast80_rag.py route \
  --db data/fast80_research_rag.sqlite3 \
  --stage experiment \
  --text "小样本条件下如何设计可靠的消融实验？"

启动 MCP 服务器：

python -B skill/research-mentor-workflows/scripts/research_mentor_mcp.py

完整的 Codex 配置见 docs/codex-install.md。架构与检索规则见 docs/architecture.md，真实运行示例见 docs/examples.md。

真实运行示例

下面两张图由 scripts/generate_examples.py 从仓库内 SQLite 数据库和 MCP 服务器现场生成，原始 JSON 一并保存在 docs/examples/。

实验阶段多头检索

MCP 运行时与 6 个工具

Experiment-stage RAG route

MCP runtime and tools

数据文件与 Git LFS

data/fast80_research_rag.sqlite3 约 56 MB，仓库已用 .gitattributes 将 SQLite 文件交给 Git LFS：

git lfs install
git add .
git commit -m "Initial release"

原始论文、仓库快照和补充材料不在本仓库中。数据库保存的是机器抽取的研究经验记录及其状态，不等于重新分发原始材料。

许可证与重要边界

  • 软件代码使用 Apache License 2.0。

  • data/ 下的派生 SQLite 数据库与 manifest 使用 CC BY 4.0。该许可不覆盖原始论文、图表、补充材料、仓库或其他第三方内容。

  • PROVISIONAL_FAST80 可用于导航、找相似方法和规划下一步，不应被写成已由人工或最终审计确认的科学事实。

  • 实时前沿搜索返回发现候选。正式声称“首创”或“研究空白”前，仍需核验全文、时间范围和最近邻工作。

  • 系统不会替代作者对原始数据、伦理、许可、引文、署名和投稿声明的责任。

仓库结构

data/                       FAST80 SQLite RAG 与 manifest
skill/research-mentor-workflows/
                            Skill、MCP、路由器、参考规范与测试
scripts/smoke_test.py       一次本地冒烟测试
scripts/install_codex.py    安装预检与配置片段生成器
tests/test_release.py       发布包结构与功能测试
docs/                       安装、架构与项目简介

项目简介

短版介绍在 docs/project-intro.zh-CN.md。

Available Tools

6 tools
research_bootstrapA

Load the compact 1200-paper RAG runtime and L0-L5 workflow status.

ParametersJSON Schema
NameRequiredDescriptionDefault

No parameters

TDQS

A3.7/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden of behavioral disclosure. It states what it does ('Load the compact 1200-paper RAG runtime and L0-L5 workflow status') but does not explain side effects, resource intensity, whether it is idempotent, or what the output looks like. This is a significant gap for a tool that likely has operational implications.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, front-loaded sentence with no filler. Every word adds value, precisely conveying the tool's purpose in a compact form.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness3/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's low complexity (no parameters, no output schema), the description is somewhat adequate but leaves gaps. It does not explain what 'L0-L5 workflow status' means or how the agent should use this status. It also does not provide context on how this bootstrapping fits with sibling tools. While not incomplete, it would benefit from a brief note on when to call it or what to expect.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The input schema is empty with zero parameters, so there is nothing that needs explanation. The description adds no parameter information, but none is required. The baseline for zero parameters is 4, as the description does not need to compensate for missing schema fields.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description uses a specific verb ('Load') and specific resources ('compact 1200-paper RAG runtime' and 'L0-L5 workflow status'), making its function clear. It also distinguishes itself from sibling tools like research_rag_query and research_route, which are query/route operations, whereas bootstrap is an initialization action.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description implies this tool is for bootstrapping or setup, but it does not explicitly state when to use it relative to alternatives. It suggests it might be called before other research operations, but there is no explicit 'use this before querying' guidance or mention of when not to use it.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

research_controllerC

Auto-detect a research stage, run mentor RAG, and require live frontier search for novelty before answering.

ParametersJSON Schema
NameRequiredDescriptionDefault
limitNo
stageNo
promptYes

TDQS

C2.7/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Since no annotations are provided, the description must carry the full burden of disclosing behavior. It reveals the mandatory live frontier search step, which is useful. However, it doesn't explain what 'run mentor RAG' entails, whether there are side effects, rate limits, or what the output looks like. This leaves significant behavioral uncertainty.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

A single, efficient sentence that packs several key behaviors without fluff. It is concise and front-loaded, though the dense jargon could be clearer.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

As a controller tool that orchestrates other research tools, the description lacks a holistic explanation of the workflow, its relationship to sibling tools, and expected outputs. With no output schema and minimal behavioral disclosure, it's insufficient for an agent to fully understand when and how to invoke it.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters2/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, so the description must compensate. It only alludes to research stage auto-detection, which hints at the 'stage' parameter, but it doesn't clarify the enum values, nor does it address 'prompt' or 'limit' at all. The 'limit' parameter is entirely unexplained.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description states a specific composite action: auto-detecting a research stage, running mentor RAG, and requiring live frontier search. This goes beyond a tautology and gives a clear sense of the tool's role. It distinguishes itself from siblings by being a controller/orchestrator, though it doesn't explicitly mention them.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No explicit guidance on when to use this tool versus alternatives. The description implies it's for a full research pipeline, but it doesn't say 'use this for complete workflows' or contrast with simpler tools like research_route. An agent would have to infer usage context.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

research_mentorD

Combine stage RAG retrieval with 12-Skill specialist method selection.

ParametersJSON Schema
NameRequiredDescriptionDefault
limitNo
stageYes
decisionYes

TDQS

D1.5/5.0
Behavior1/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description carries the full burden for behavioral disclosure. It does not mention side effects, return values, prerequisites, or any operational traits beyond the vague 'combine' statement, leaving the agent without critical safety or behavior information.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness2/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single short sentence, but it is under-specified rather than concise. It omits essential information, making the brevity a liability rather than an asset.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness1/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's moderate complexity (3 parameters, required fields, enum) and the absence of an output schema and annotations, the description is severely incomplete. It does not provide enough context for an agent to correctly invoke the tool or interpret its results.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0% and the description does not explain any of the three parameters (stage, decision, limit). Although 'stage' appears in the phrase 'stage RAG retrieval', it does not clarify the parameter's meaning, format, or how it relates to the decision parameter.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose2/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description states 'Combine stage RAG retrieval with 12-Skill specialist method selection' but lacks a clear verb+resource structure. It does not specify what the tool actually does with these components or what output to expect, making it vague and not easily distinguishable from siblings.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No guidance is provided on when to use this tool versus alternatives. While sibling tools like research_rag_query and research_route exist, the description does not mention any context, exclusions, or alternative selection criteria.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

research_rag_queryC

Query the local embedding-free 1200-paper RAG.

ParametersJSON Schema
NameRequiredDescriptionDefault
kindNo
textYes
limitNo

TDQS

C2.7/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries full responsibility for disclosing behavior. It mentions 'embedding-free' and 'local', but doesn't state whether the operation is read-only, what the query does exactly, or what limitations or side effects exist.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single, focused sentence with no wasted words, front-loading the core action and resource. However, it may be slightly too terse, sacrificing important context for brevity.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Without an output schema or annotations, the description must explain what the query returns and how parameters influence the result. It doesn't, and the sibling tool list doesn't compensate for this missing context. The description is too minimal for a tool with three parameters and no output schema.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters2/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, and the description adds no explanations for the parameters 'text', 'kind', or 'limit'. The agent must guess what 'text' should contain and what 'kind' controls, leaving a significant semantic gap.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's action ('Query') and resource ('local embedding-free 1200-paper RAG'), making the purpose unmistakable. It doesn't explicitly differentiate from sibling tools, but the resource specification is distinct enough to separate it from general research tools.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

There is no guidance on when to use this tool versus alternatives like research_frontier_search or research_controller. The description only states what it does, not the scenario in which it should be selected.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

research_routeC

Run mandatory multi-head stage retrieval and return next action/correction.

ParametersJSON Schema
NameRequiredDescriptionDefault
limitNo
stageYes
decisionYes

TDQS

C2.2/5.0
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations provided, the description is solely responsible for disclosing behavior. It states that a retrieval runs and returns a result, but it doesn't mention permissions, side effects, reversibility, or what 'correction' entails. This leaves critical behavioral aspects undisclosed.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is a single sentence, which is brief and front-loaded with a verb. However, it uses unexplained jargon ('multi-head stage retrieval') and lacks substance, making it under-specified rather than appropriately concise. It is short but not informative.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool has 3 parameters, no output schema, and no annotations, the description should do heavy lifting. It fails to clarify what the parameters mean, what the returned 'action/correction' looks like, or how the tool fits into a workflow. Major gaps remain in even basic understanding.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters1/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 0%, so the description must compensate by explaining the parameters. It provides no information about 'stage', 'decision', or 'limit' — not even a hint of how they influence the tool's behavior. The agent cannot infer what values to provide or why.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose3/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description identifies a verb ('Run') and a resource ('multi-head stage retrieval'), and mentions a return value ('next action/correction'). However, 'multi-head stage retrieval' is jargon that isn't explained, and it doesn't clearly differentiate from sibling tools like research_controller or research_mentor. The overall purpose is only loosely conveyed.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No guidance is given about when to use this tool versus alternatives. The word 'mandatory' hints at a required step but doesn't clarify the context or exclusions. There is no mention of alternatives or decision criteria.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

TDQS

C2.8/5.0
Disambiguation4/5

Each tool targets a distinct aspect of the research workflow, but research_route and research_mentor both involve stage retrieval. Their descriptions clarify different purposes: route returns next actions while mentor selects specialist methods, so most agents can differentiate them.

Naming Consistency4/5

All tools share the consistent 'research_' prefix, and most use action-oriented names like bootstrap, route, and mentor. The mix of verb and noun forms (e.g., controller vs. frontier_search) is a minor deviation from a strict verb_noun pattern.

Tool Count5/5

Six tools are well-scoped for a research mentor RAG system, covering initialization, direct query, routing, mentoring, live search, and orchestration. This is a reasonable number that avoids both thinness and bloat.

Completeness4/5

The tool set covers the full workflow from bootstrapping the RAG runtime to answering after a novelty search, including a controller that orchestrates the process. Minor gaps exist, such as no explicit tool for updating the RAG corpus, but core operations are complete.

Maintenance

ActivityMaintained
ResponsivenessSyncing

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/GelortZCY/research-mentor-rag'

If you have feedback or need assistance with the MCP directory API, please join our Discord server