# Ghostlight

> Use this tool when you need to verify the accuracy of AI-generated responses and prevent hallucination. Ghostlight solves problems related to accessibility, visual regression, and UI verification through adversarial verification with screenshots and subagents. It takes in AI outputs and returns verified results, making it ideal for use cases where reliability and trustworthiness of AI responses are crucial.

Canonical page: https://skillsregistry.net/skills/banditburai-ghostlight  
JSON: https://api.skillsregistry.net/v1/skills/banditburai-ghostlight

## Description

A QA verification MCP server that prevents AI hallucination by using adversarial verification with screenshots and subagents for accessibility, visual regression, and UI verification.

## Trust

- **Trust score (0–1):** 0.70
- **Verification tier:** verified
- **Last scanned:** 2026-08-31

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** other
- **Updated:** 2026-08-31

## Source

- **Source listing:** [Glama](https://glama.ai/mcp/servers/c5tffjcy2p)
- **Repository:** <https://github.com/banditburai/ghostlight>

## 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": "banditburai-ghostlight"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/banditburai-ghostlight` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/banditburai-ghostlight/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