Skip to main content
Glama

OpsLevel MCP

Official
README.md3.62 kB
# mcp-eval Evaluate result of various queries/prompts using opslevel-mcp. At the moment is just uses Anthropic's Claude API, but OpenAI support should come soon. ## Setup Copy the `template.env` to `.env` and set the necessary variables. ```sh ANTHROPIC_API_KEY='' # required OPSLEVEL_APP_URL='' # optional OPSLEVEL_API_TOKEN='' # required MCP_SERVER_PATH='' # required ``` You can also pass these as flags to the command line. See `yarn eval --help` for more details ## Prompts Prompts are kept in `prompts.js`. Each prompt has a `slug` and a `query`. ## Running The `MCP_SERVER_PATH` environment variable is the path on your local file system that points to the MCP server you want to test. ``` yarn run eval ``` If you don't set the `MCP_SERVER_PATH` environment variable, you must provide the path as a command-line argument. ``` yarn run eval path/to/opslevel-mcp ``` If you provide both the `MCP_SERVER_PATH` variable and the command-line argument, the command-line argument takes precedence. For more detailed logging: ``` DEBUG=true yarn run eval ``` ## Results The results are in the `results` folder. Each run creates a new folder with the time of the run (eg: `2025-04-21T17:57/`). For each slug in `prompts.js`, there's a `<slug>.json` file. Example: ```json { "prompt": { "slug": "employees", "query": "Who works at opslevel" }, "response": "Based on the information retrieved from the OpsLevel account, there are 2 users registered:\n\n1. **Alice**\n - Email: alice@opslevel.com\n - Role: Admin\n\n2. **Foobar**\n - Email: foobar@example.com\n - Role: Team Member\n\nThese are the individuals who have accounts in this OpsLevel system. Alice appears to have administrator privileges, while Foobar is a regular team member.", "raw_messages": [ { "role": "user", "content": "Who works at opslevel" }, { "role": "assistant", "content": [ { "type": "text", "text": "I'll help you find information about who works at OpsLevel. Let me retrieve the list of users in the OpsLevel account." }, { "type": "tool_use", "id": "toolu_013HeznQfrB3tLQbRaDfvPdi", "name": "users", "input": {} } ] }, { "role": "user", "content": [ { "type": "tool_result", "tool_use_id": "toolu_013HeznQfrB3tLQbRaDfvPdi", "content": [ { "type": "text", "text": "[{\"Id\":\"Z2lkOi8vb3BzbGV2ZWwvVXNlci8x\",\"Email\":\"alice@opslevel.com\",\"HTMLUrl\":\"http://app.opslevel.local:3000/users/1\",\"Name\":\"Alice\",\"Role\":\"admin\"},{\"Id\":\"Z2lkOi8vb3BzbGV2ZWwvVXNlci8y\",\"Email\":\"foobar@example.com\",\"HTMLUrl\":\"http://app.opslevel.local:3000/users/2\",\"Name\":\"Foobar\",\"Role\":\"team_member\"}]" } ] } ] }, { "role": "assistant", "content": [ { "type": "text", "text": "Based on the information retrieved from the OpsLevel account, there are 2 users registered:\n\n1. **Alice**\n - Email: alice@opslevel.com\n - Role: Admin\n\n2. **Foobar**\n - Email: foobar@example.com\n - Role: Team Member\n\nThese are the individuals who have accounts in this OpsLevel system. Alice appears to have administrator privileges, while Foobar is a regular team member." } ] } ], "run_at": "2025-04-21T17:57:49.157Z", "ops_level_mcp_version": "unknown", "model_config": { "model": "claude-3-7-sonnet-20250219", "max_tokens": 1000 } } ```

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/OpsLevel/opslevel-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server