Content
# Tool List
> AI workflow project for PE / VC / FA / industrial investors, currently focusing on organizing Prompt, Skill, Workflow, and providing MCP call capabilities.
## Official Website
**Production Site: [https://nei-pevc.com](https://nei-pevc.com)**
[Access Product](https://nei-pevc.com) · [MCP Configuration Guide](https://nei-pevc.com/mcp) · [Security and Confidentiality Principles](https://nei-pevc.com/security) · [Contribute](./CONTRIBUTING.md)
This repository is the main repository for `nei-pevc.com`, which is also the code repository currently used by the online site.
## About N.E.I.
**N.E.I. = New Era Investor**.
N.E.I. is a next-generation investor community initiated by jammyssy, currently with over 500 members, focusing on the primary market, AI tools, investment workflow, and institutional efficiency improvement.
`nei-pevc.com` is a PEVC AI product under the N.E.I. community, originating from specific frontline work issues:
- Too many BPs, can we have a reliable initial screening framework?
- Can we reduce repetitive work in research, due diligence, IC, and post-investment stages?
- Can we reuse good Prompts, templates, and skills in one place?
- Can we directly call skills from Claude Code, Cursor, and Windsurf?
Currently, the project focuses on organizing commonly used Prompts, Skills, Workflows, and templates for the primary market, supporting search, collection, and replication, and providing AI clients with call capabilities through MCP.
You can think of it as a combination of three things:
- A Prompt/Skill library that primary market practitioners can understand;
- A PEVC MCP Server that can connect to Claude Code, Cursor, and Windsurf;
- A place to accumulate institutional methodologies and work templates.
The project is currently in the small-scale public beta stage, and functions, content, and collaboration methods will continue to be adjusted.
## Why Build This Site
Primary market work involves many judgment processes that seem different on the surface but have highly repetitive underlying methods:
- Should a BP enter the initial screening?
- How to quickly form an investment map for a track?
- What questions should be asked in commercial due diligence interviews?
- How to structure IC Memo and counterparty questions?
- How to standardize post-investment reports, government responses, and LP reports?
These methods are often scattered in personal experiences, Word templates, Excel lists, investment committee materials, and chat records. Over time, each person develops their own processing method, but it's hard to reuse and pass on to AI.
`nei-pevc.com` starts with reusable content: putting usable Skills, Prompts, Workflows, and templates together to reduce repetitive sorting and writing in real work.
## What Can Be Done Now
### 1. Skill Library
Browse, search, and filter Prompts, SKILL.md, templates, methodologies, tool combinations, and Workflows in PEVC scenarios.
Supported dimensions include:
- Work scenarios: BP initial screening, industry research, commercial due diligence, financial analysis, IC materials, post-investment management, fundraising/LP, etc.
- Content types: Prompt, SKILL.md, Workflow, Template, Tool Stack, Case Study
- Industry tracks: AI, semiconductors, new energy, medical health, consumer, advanced manufacturing, etc.
- Work content: information collection, due diligence questions, report generation, automation, risk identification, etc.
### 2. Favorites Library
After logging in, you can collect commonly used Skills in your Library and add notes, sorting, and reuse clues for each collection.
Favorites will become the basis for MCP calls. Later, your AI client can read your Favorites Library through `list_my_skills`, reducing the cost of re-searching from the entire site.
### 3. MCP Server
The site provides an MCP Server. After configuring the Token, you can search, read, and call Skills on the AI client.
Current tools include:
- `search_skills`: search public Skills by keyword/scene/type
- `get_skill`: get the complete Prompt/Workflow text of a Skill
- `list_my_skills`: list Skills that the current user has collected and approved for MCP
- `apply_skill`: fill in the context into the Prompt template and return the executable Prompt
- `favorite_skill`: add public Skills to the Favorites Library from the client
Typical usage:
```text
Use N.E.I. to help me initially screen this BP and determine if it should be established.
Use N.E.I. to generate a commercial due diligence interview question list.
Use N.E.I. to write a primary structure and key risks for an IC Memo.
Use N.E.I. to draft a formal reply letter to a government guidance fund.
```
### 4. Skill Quality System
The details page will display Skill Quality, applicable scenarios, input examples, output expectations, and suggested supplements.
The rating mainly determines the reuse degree of the content: which content is already relatively complete, and which needs further supplementation. It will consider:
- Is the title specific?
- Are the scene and tags complete?
- Is the text sufficiently structured?
- Does it contain placeholders, steps, input/output examples?
- Does the file-type Skill have attachments?
- Is there a source, installation instructions, and usage experience?
The homepage will prioritize content that is truly formed and has a high-quality score.
### 5. Admin Operation Platform
The administrator's side currently mainly handles:
- Pending review queue
- Reported content
- MCP approval
- Homepage selection
- Error monitoring
- API/MCP call data
These operational capabilities will determine the order after the content scale expands. As the content increases, review, reporting, MCP approval, and error monitoring are necessary to maintain quality.
## MCP Security and Confidentiality Principles
This site will be used to handle investment work, so the security boundary must be clear and not ambiguous.
N.E.I. MCP defaults:
- Do not read local files
- Do not upload user files
- Do not save BP, financial models, investment committee materials, LP lists
- Only distribute Skill/Workflow text
- Token can be revoked and regenerated at any time
- User-submitted Skills are reviewed before entering MCP
- Call logs do not record user-sensitive text
See [MCP Security and Confidentiality Principles](https://nei-pevc.com/security) for a complete description.
## Tech Stack
- **Next.js 15 App Router**
- **React 18**
- **TypeScript**
- **Tailwind CSS**
- **Prisma 5**
- **PostgreSQL / Neon**
- **MCP SDK + mcp-handler**
- **TipTap Rich Editor**
- **Vercel deployment**
## Local Development
```bash
npm install
npm run db:push
npm run dev
```
Access:
```text
http://localhost:3000
```
Common commands:
```bash
npm run lint
npx tsc --noEmit --pretty false
npm run smoke:public-posts
npm run build
```
Production build:
```bash
npm run build
```
Vercel build:
```bash
npm run vercel-build
```
## Environment Variables
Refer to [.env.example](./.env.example).
Key variables include:
- `DATABASE_URL`
- `SESSION_SECRET`
- `NEXT_PUBLIC_BASE_URL`
- Email service configuration
- AI/GLM/Anthropic-related configuration
- File storage-related configuration
Do not submit `.env` or `.env.local`.
## Project Structure
```text
app/ Next.js App Router pages and API
components/ UI, homepage modules, MCP components, detail page components
lib/ database, session, security, MCP, formatting, business logic
prisma/ Prisma schema and migration
scripts/ data import, migration, online smoke test
uploads/ local development attachment directory
```
Key directories:
- `app/api/mcp/route.ts`: MCP Server entry
- `app/connect/page.tsx`: MCP onboarding and Token configuration
- `app/security/page.tsx`: security and confidentiality principles
- `app/admin/page.tsx`: Admin operation platform
- `components/home/`: homepage product modules
- `lib/mcp-safety.ts`: MCP return content safety prefix
- `lib/skill-quality.ts`: Skill quality scoring
- `scripts/smoke-public-posts.ts`: public post detail page smoke test
## Pre-Launch Checklist
Before each release, it is recommended to run at least:
```bash
npm run lint
npx tsc --noEmit --pretty false
npm run smoke:public-posts
npm run build
```
If you want to check online detail pages:
```bash
npm run smoke:public-posts -- --base https://nei-pevc.com
```
## How to Participate
Two types of contributions are welcome:
- **Content contribution**: submit reusable Prompts, SKILL.md, Workflows, templates, cases, and tool combinations in PEVC scenarios.
- **Product/code contribution**: fix bugs, optimize pages, supplement MCP capabilities, and improve security and operation tools.
Even if you don't have an engineering background, you can directly post on the website or write a real work scenario in GitHub Issues. For example, "I want to use AI to help me initially screen BP, but I don't know how to ask," this demand itself is very valuable. Maintainers will then organize it into Skills, Workflows, or product demands.
See [CONTRIBUTING.md](./CONTRIBUTING.md) for complete contribution methods.
## Current Progress
The project is currently in the small-scale public beta stage, suitable for:
- Finding 10-30 PEVC/FA/industrial investment friends for small-scale trial use;
- Collecting real tasks, reducing empty design functions;
- Continuing to supplement BP initial screening, IC materials, post-investment management, and other shortcomings;
- Verifying whether MCP Favorites Library can really save time.
The next stage will not focus solely on "adding more content." More importantly:
- Improve core scenario coverage
- Enhance security trust expression
- Improve Skill quality threshold
- Make task-to-Skill Bundle recommendation paths
- Precipitate institutional-level private Skill Repository
## Project Origin and Maintenance
N.E.I. (New Era Investor) is a next-generation investor community initiated and organized by jammyssy with 500+ members. `nei-pevc.com` is a PEVC-focused AI product line under the broader N.E.I. community. The earliest MVP was scaffolded with support from jammyssy. The `nei-pevc.com` product line — including product direction, MCP workflow, security system, content model, admin operations, deployment, infrastructure, domain, and ongoing development — is led, funded, and maintained by lensnowovo. This repository is the canonical product repository for `nei-pevc.com`.
See also: [NOTICE.md](./NOTICE.md).
## License
Private beta project. All rights reserved unless otherwise stated.
MCP Config
Below is the configuration for this MCP Server. You can copy it directly to Cursor or other MCP clients.
mcp.json
Connection Info
You Might Also Like
No recommendations available