# genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill

> genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill — alphaparkinc-genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill. Use this tool when you need to evaluate and improve the performance of language models using human preference rubrics and red-teaming pipelines. It solves problems related to model alignment and safety by identifying potential flaws and biases through rigorous testing and feedback. The tool takes in model outputs and human evaluations as inputs and produces refined model performance and identified weaknesses as outputs.

Canonical page: https://skillsregistry.net/skills/alphaparkinc-genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill  
JSON: https://api.skillsregistry.net/v1/skills/alphaparkinc-genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill

## Description

RLHF human preference rubric and DPO red-teaming pipeline (Surge AI style)

## Trust

- **Trust score (0–1):** 0.50
- **Verification tier:** unverified

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** container
- **Runtime environment:** vm
- **Updated:** 2026-09-22

## Source

- **Source listing:** [GitHub](https://github.com/alphaparkinc/genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill)

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "alphaparkinc-genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/alphaparkinc-genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/alphaparkinc-genpark-rlhf-human-preference-rubric-red-teaming-pipeline-skill/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt