# orchescope

Source: https://glimind.com/server/npm%2Forchescope
Registry id: `npm/orchescope`
Status: unknown

## Install

```json
{
  "mcpServers": {
    "orchescope": {
      "command": "npx",
      "args": [
        "-y",
        "orchescope"
      ]
    }
  }
}
```

Claude Code:

```sh
claude mcp add orchescope -- npx -y orchescope
```

## Tools (13)

- `audit`: Discover the system, reconcile it against stored runs, report findings
- `benchmark`: Vary one dimension and compare the variants
- `chaos`: Inject the faults a scenario declares and report what each one did
- `compare`: Compare two runs, scans or revisions
- `doctor`: Check that this machine can run every command this build offers
- `export`: Export the report
- `goal_create`: Turn a finding into a bounded improvement goal
- `goal_validate`: Judge a goal against measured evidence
- `init`: Create .orchescope with a configuration file listing every default
- `mcp_serve`: Speak the Model Context Protocol on stdio (primary agent surface)
- `receive`: Listen for spans from a system that is already running, store them as a run
- `test`: Run a scenario and evaluate it
- `trace`: Run a command, collect its OpenTelemetry spans, store them as a run

Measured by Glimind (https://glimind.com): MCP handshake + tool list only, no tool is ever called. CC BY 4.0.