deepseek-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| DEEPSEEK_MODEL | No | Default model for chat. | deepseek-chat |
| DEEPSEEK_API_KEY | Yes | Required. DeepSeek API key. | |
| DEEPSEEK_BASE_URL | No | API base URL. | https://api.deepseek.com |
| HARNESS_CMD_CODEX | No | Headless command for codex host, e.g. 'codex exec --skip-git-repo-check'. | |
| HARNESS_CMD_CLAUDE | No | Headless command for claude host, e.g. 'claude -p'. | |
| HARNESS_ALLOW_HOSTS | No | Comma-separated hosts the bridge may drive. Empty = locked. | |
| HARNESS_CMD_OPENCODE | No | Headless command for opencode host, e.g. 'script -q /dev/null opencode run'. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| chatA | Send a chat completion request to DeepSeek. Accepts an OpenAI-style message list and returns the assistant's reply text. |
| list_modelsA | List models available on the DeepSeek API. |
| list_harnessesA | List harness hosts the bridge can drive (from HARNESS_ALLOW_HOSTS / HARNESS_CMD_* config) with their commands. |
| use_harnessA | Run a prompt in a headless host CLI (claude -p, codex exec, opencode run, ...). Lets the DeepSeek model drive local agents, skills, plugins, and MCP tools configured on that host. Host must be in HARNESS_ALLOW_HOSTS. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 4 tools
chat and use_harness both send prompts, but one targets the DeepSeek API and the other targets local harness CLIs, which is clearly distinguished in descriptions. list_models and list_harnesses are distinct list operations for different resources.
All tool names follow a clear verb style: chat, list_models, list_harnesses, use_harness. The naming is consistent and predictable, with list_* for listing operations and direct verbs for actions.
Four tools is well-scoped for a bridge server that provides chat completion, model listing, and harness control. Each tool serves a distinct purpose without redundancy.
The core workflow of sending prompts and listing models/harnesses is covered. A minor gap is lack of streaming or more detailed model/harness management, but the essential lifecycle for the stated purpose is present.