careerwin-mcp
README.md
# careerwin-mcp
[](./LICENSE)
[](https://modelcontextprotocol.io)
[](./package.json)
[](https://careerwin.ai)
**Model Context Protocol (MCP) Server for O\*NET 31.0 / ESCO v1.2.1 7-Primitive Occupation Resolution, 4-Gatekeeper ATS Verification, and 4-Seat Adversarial Hiring Committee Simulation.**
> Connect Claude Desktop, Cursor, Windsurf, or Google Antigravity directly to the deterministic career & ATS verification primitives powering **[CareerWin (`careerwin.ai`)](https://careerwin.ai)** and **[`Richard-Ewing/careerwin`](https://github.com/Richard-Ewing/careerwin)**.

---
## Why Generic LLM Resume Prompts Fail Real Hiring Committees
Ask a standard LLM to evaluate or rewrite a resume for a **Trauma ICU Registered Nurse**, a **Commercial Master Electrician**, or a **Staff Distributed Systems Engineer**, and it makes the same three mistakes:
1. **Tech / Product Management Hallucination Bias**: It injects SaaS buzzwords (*"cross-functional stakeholder alignment"*, *"product roadmap synergy"*) into clinical, skilled trades, and finance records.
2. **Statutory Gate Blindness**: It treats non-substitutable legal credentials (`Active State RN License`, `Master Electrician / Red Seal`, `CPA`, `Active TS/SCI`) as optional soft skills that can be bridged by transferable activities.
3. **AI Cliché Contamination**: It fills bullets with *spearheaded, leveraged, utilized, dynamic, passionate, seamless, robust, synergy*, triggering immediate skepticism from human hiring managers.
`careerwin-mcp` gives your AI coding or desktop agent 3 deterministic tools grounded in **O\*NET 31.0 + ESCO v1.2.1** across all 23 major Standard Occupational Classification (SOC) families.
---
## Exposed MCP Tools
| Tool Name | Purpose | Output |
| :--- | :--- | :--- |
| `resolve_occupation_primitives` | Decomposes any job title into 7 Work Primitives (`ACTIVITIES`, `DECISIONS`, `TOOLS`, `CONSTRAINTS`, `SCOPE`, `OUTCOMES`, `NON_SUBSTITUTABLE_STATUTORY_GATES`). | `ONET_31_0_ESCO_1_2_1` schema + resolution mode (`KNOWN_OCCUPATION_ADAPTER`, `DYNAMIC_SYNTHESIS`, `UNRESOLVED_OCCUPATION_MODEL`). |
| `audit_ats_gatekeepers` | Evaluates a resume against the **4 Gatekeepers** (ATS Statutory Parser, Recruiter Boolean Search, Hiring Manager Method & Scope, Bar Raiser Calibration). | Composite score (`0-100`), missing statutory gates, quantified evidence count, and banned AI cliché hits. |
| `simulate_hiring_committee` | Runs a **4-Seat Behind-Closed-Doors Adversarial Hiring Committee** (`Recruiter Gatekeeper`, `Hiring Manager`, `Bar Raiser`, `Compensation Chair`). | 4-Field Seat Contracts (`[ANCHORED]` / `[INFERRED]` / `[SPECULATIVE]`), `FLIP:` triggers, and a 3-Form Ruling (`APPLY_NOW`, `APPLY_ONLY_IF`, `DO_NOT_APPLY`). |
---
## 1-Minute Setup (Claude Desktop / Cursor / Antigravity)
Add `careerwin-mcp` to your `claude_desktop_config.json` or `.cursor/mcp.json`:
```json
{
"mcpServers": {
"careerwin": {
"command": "npx",
"args": ["-y", "careerwin-mcp"]
}
}
}
```
Or test the tools interactively from your terminal right now:
```bash
npx careerwin-mcp --inspect
```
---
## Architecture
```mermaid
flowchart LR
A["Claude Desktop / Cursor / Antigravity"] <-->|JSON-RPC 2.0 Stdio| B["careerwin-mcp Server"]
B --> C["resolve_occupation_primitives (O*NET 31.0 / ESCO v1.2.1)"]
B --> D["audit_ats_gatekeepers (4-Gatekeeper Verifier + Cliché Detector)"]
B --> E["simulate_hiring_committee (4-Seat Adversarial Council)"]
C & D & E --> F["Full 5-Surface Opportunity Kits & 380ms ATS Autofill at https://careerwin.ai"]
```
---
## Upgrade to the Full CareerWin Platform
Need the full cloud compiler with:
- **5-Surface Opportunity Kits** (Single-Column ATS Vector PDF, True OOXML `.docx`, Executive Cover Letter, Hiring Manager Cold Outreach, and Screener Q&A)
- **12-Dimension 4-Gatekeeper LinkedIn & Resume Studio**
- **Chrome Extension v5.1.83 (`shadow-ats-piercer`)** for 1-click 380ms Workday, Greenhouse, Lever, and Ashby autofill
Launch **[CareerWin (`https://careerwin.ai`)](https://careerwin.ai)** or explore **[`Richard-Ewing/careerwin`](https://github.com/Richard-Ewing/careerwin)**.
---
## License
MIT © [Richard Ewing](https://github.com/Richard-Ewing) & [CareerWin AI](https://careerwin.ai)
This server cannot be deployed
Maintenance
ActivityMaintained
ResponsivenessNo issues