> ## Documentation Index
> Fetch the complete documentation index at: https://docs.budecosystem.com/llms.txt
> Use this file to discover all available pages before exploring further.

# MCP Foundry Reference

> This reference maps common platform tasks to the dashboard modules and the MCP concepts exposed to clients.

## Dashboard modules

| Module                    | Use it to                                                            | Key outcome                                                       |
| ------------------------- | -------------------------------------------------------------------- | ----------------------------------------------------------------- |
| Dashboard                 | Review operational activity and health.                              | Visibility into traffic, latency, errors, and execution activity. |
| Virtual Servers           | Compose and publish a curated MCP endpoint.                          | A client-facing server with selected capabilities.                |
| MCP Servers               | Connect and manage external, self-hosted, or deployable MCP servers. | Federated or managed capability sources.                          |
| MCP Registry              | Discover available pre-integrated servers.                           | Faster onboarding for common services.                            |
| Playground                | Interact with a selected server.                                     | Functional validation before and after release.                   |
| Tools                     | Define and manage direct API-backed operations.                      | Agent-callable functions with schemas and authentication.         |
| MCP Tools Generator       | Generate candidate tools from API sources.                           | Reduced manual API-to-tool work.                                  |
| Generation Jobs           | Track tool-generation work.                                          | Reviewable job progress and outcomes.                             |
| Resources                 | Add URL-based or inline agent context.                               | Discoverable organizational knowledge.                            |
| Prompts                   | Create reusable, parameterized templates.                            | Consistent agent instructions.                                    |
| Skills Hub                | Author and manage portable markdown procedures.                      | Reusable agent capabilities tied to tools.                        |
| API Tokens                | Create and manage programmatic credentials.                          | Controlled application access.                                    |
| Roots                     | Define permitted filesystem directories.                             | Constrained server filesystem access.                             |
| Logs                      | Inspect detailed activity records.                                   | Traceability and troubleshooting.                                 |
| Teams and User Management | Manage membership, ownership, and access.                            | Governed collaboration.                                           |

## Capability fields

The exact form fields can vary by capability and deployment configuration, but use this checklist when reviewing an entry.

| Field or concern              | Tools                             | Resources                 | Prompts                     | Skills                                   | Virtual servers                             |
| ----------------------------- | --------------------------------- | ------------------------- | --------------------------- | ---------------------------------------- | ------------------------------------------- |
| Name and description          | Yes                               | Yes                       | Yes                         | Yes                                      | Yes                                         |
| Tags                          | Yes                               | Yes                       | Yes                         | Optional                                 | Optional                                    |
| Ownership and visibility      | Yes                               | Yes                       | Yes                         | Yes                                      | Yes                                         |
| Backend connection or content | Endpoint, headers, authentication | URL or inline content     | Template and arguments      | Markdown instructions and selected tools | Client connection and selected capabilities |
| Input contract                | JSON Schema                       | Resource URI/content type | Declared template arguments | Procedure and expected inputs            | Exposed catalog selection                   |

## MCP client interaction

Once a client is authorized to connect to a published virtual server, it uses MCP discovery and invocation methods to work with the server's exposed catalog.

```mermaid theme={null}
sequenceDiagram
    participant Client as MCP client
    participant Server as Virtual server
    participant Capability as Attached capability
    Client->>Server: initialize
    Client->>Server: tools/list, resources/list, prompts/list
    Server-->>Client: Curated catalog
    Client->>Server: tools/call or resource/prompt request
    Server->>Capability: Validate, authorize, and execute
    Capability-->>Server: Result
    Server-->>Client: MCP response
```

| MCP interaction  | Purpose in MCP Foundry                                                   |
| ---------------- | ------------------------------------------------------------------------ |
| `initialize`     | Establishes an MCP session and negotiates client/server capabilities.    |
| `tools/list`     | Lists tool entries exposed by the selected virtual server.               |
| `tools/call`     | Invokes an exposed tool with arguments validated against its definition. |
| `resources/list` | Lists resources available through the selected virtual server.           |
| `resources/read` | Retrieves content for an exposed resource.                               |
| `prompts/list`   | Lists prompt templates available through the selected virtual server.    |
| `prompts/get`    | Retrieves a named prompt with its supplied arguments.                    |

## Common configuration patterns

### Read-only knowledge assistant

Attach read-only search or retrieval tools, policy resources, and a response-format prompt to a narrowly scoped virtual server. Do not attach mutation tools.

### Action-taking service assistant

Attach the least-privileged tools required for the workflow, a skill with approval and escalation rules, resources containing policy context, and an audit-friendly team ownership model.

### Federated enterprise catalog

Connect approved MCP servers, review the capabilities they provide, and create role- or domain-specific virtual servers instead of exposing every imported capability through one endpoint.

## Related pages

* [Concepts](https://docs.budecosystem.com/mcp-foundry/mcp-foundry-concepts) explains composition, federation, and governance.
* [Configure MCP Server](https://docs.budecosystem.com/mcp-foundry/guides/configuring-gateways) provides step-by-step configuration workflows.
* [Troubleshooting](https://docs.budecosystem.com/mcp-foundry/troubleshooting) covers common failure modes.