# genpark-constitutional-ai-harmlessness-evaluator-benchmark-skill

> genpark-constitutional-ai-harmlessness-evaluator-benchmark-skill — alphaparkinc-genpark-constitutional-ai-harmlessness-evaluator-benchmark-skill. Use this tool when you need to evaluate the harmlessness of AI models and ensure their safety, particularly in relation to constitutional AI and Anthropic-style risk assessment. It provides a benchmark for measuring AI safety and offers a safety evaluator for RLAIF (Reinforcement Learning from Artificial Intelligence Feedback) systems. This tool is ideal for developers and researchers seeking to assess and mitigate potential risks associated with AI model deployment.

Canonical page: https://skillsregistry.net/skills/alphaparkinc-genpark-constitutional-ai-harmlessness-evaluator-benchmark-skill  
JSON: https://api.skillsregistry.net/v1/skills/alphaparkinc-genpark-constitutional-ai-harmlessness-evaluator-benchmark-skill

## Description

Constitutional AI harmlessness benchmark and RLAIF safety evaluator (Anthropic style)

## Trust

- **Trust score (0–1):** 1.00
- **Verification tier:** verified
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** container
- **Runtime environment:** vm
- **Updated:** 2026-09-28

## Source

- **Source listing:** [GitHub](https://github.com/alphaparkinc/genpark-constitutional-ai-harmlessness-evaluator-benchmark-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-constitutional-ai-harmlessness-evaluator-benchmark-skill"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/alphaparkinc-genpark-constitutional-ai-harmlessness-evaluator-benchmark-skill` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/alphaparkinc-genpark-constitutional-ai-harmlessness-evaluator-benchmark-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