Truesight MCP Skills

Agent skills and Cursor plugin for the Truesight MCP - step-by-step workflow playbooks for scoring inputs, building live evaluations, error analysis, and the review loop.

Works with Claude Code, Cursor, and any client that supports the agent skills standard.

Prerequisites

A Truesight account
A platform API key — go to Settings in Truesight, click Create Key, select your scopes, and copy the key

Connect the MCP

Cursor

Click the button, then replace YOUR_API_KEY_HERE with your platform API key in ~/.cursor/mcp.json.

Alternatively, this repo is a Cursor Marketplace plugin — installing it adds the Truesight MCP server and both skills to Cursor in one step (then replace YOUR_API_KEY_HERE in ~/.cursor/mcp.json).

Or add it manually to .cursor/mcp.json:

{
  "mcpServers": {
    "truesight": {
      "url": "https://api.truesight.goodeyelabs.com/mcp/",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY_HERE"
      }
    }
  }
}

VS Code (GitHub Copilot)

Click a button, then replace YOUR_API_KEY_HERE with your platform API key. Requires VS Code 1.99+ with GitHub Copilot enabled.

Or add it manually to .vscode/settings.json:

{
  "mcp": {
    "servers": {
      "truesight": {
        "url": "https://api.truesight.goodeyelabs.com/mcp/",
        "headers": {
          "Authorization": "Bearer YOUR_API_KEY_HERE"
        }
      }
    }
  }
}

Claude Code

claude mcp add --transport http truesight \
  https://api.truesight.goodeyelabs.com/mcp/ \
  --header "Authorization: Bearer YOUR_API_KEY_HERE"

Add --scope project before truesight to scope it to a single project instead of your user config.

Claude Desktop

Open Claude → Settings → Developer → Edit Config and add to claude_desktop_config.json:

{
  "mcpServers": {
    "truesight": {
      "url": "https://api.truesight.goodeyelabs.com/mcp/",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY_HERE"
      }
    }
  }
}

Save and restart Claude.

Windsurf

Open Settings → Cascade → MCP Servers → View raw config and add:

{
  "mcpServers": {
    "truesight": {
      "serverUrl": "https://api.truesight.goodeyelabs.com/mcp/",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY_HERE"
      },
      "disabled": false
    }
  }
}

Save and click Refresh (or restart Windsurf).

Skills

| Skill | What it does | |-------|-------------| | truesight-mcp | Decision tree + playbooks for scoring inputs (A), error analysis (B), and the review-and-promote loop (C) | | create-evaluation | Scope, build, and deploy a new live evaluation from scratch — includes judgment_configs schema reference and mandatory api_key storage guidance |

Install skills manually

Project-level (recommended for team workflows)

curl -fsSL https://raw.githubusercontent.com/Goodeye-Labs/truesight-mcp-skills/main/skills/truesight-mcp/SKILL.md \
  -o .claude/skills/truesight-mcp/SKILL.md --create-dirs

curl -fsSL https://raw.githubusercontent.com/Goodeye-Labs/truesight-mcp-skills/main/skills/create-evaluation/SKILL.md \
  -o .claude/skills/create-evaluation/SKILL.md --create-dirs

Global (available in all projects)

curl -fsSL https://raw.githubusercontent.com/Goodeye-Labs/truesight-mcp-skills/main/skills/truesight-mcp/SKILL.md \
  -o ~/.claude/skills/truesight-mcp/SKILL.md --create-dirs

curl -fsSL https://raw.githubusercontent.com/Goodeye-Labs/truesight-mcp-skills/main/skills/create-evaluation/SKILL.md \
  -o ~/.claude/skills/create-evaluation/SKILL.md --create-dirs

Usage

Once the MCP is connected and skills are installed, your AI assistant will automatically pick up the right skill based on what you ask:

"Score these inputs against my quality eval" — triggers truesight-mcp, Workflow A
"Analyze the errors in my dataset" — triggers truesight-mcp, Workflow B
"Review and promote these flagged results" — triggers truesight-mcp, Workflow C
"Create an evaluation for response quality" — triggers create-evaluation

License

MIT

MCP Servers