nexus-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| NEXUS_CODEX_MODEL | No | Default model for the Codex runner | |
| NEXUS_CLAUDE_MODEL | No | Default model for the Claude runner | |
| NEXUS_CODEX_MODELS | No | Comma-separated model list for the Codex runner | |
| NEXUS_GEMINI_MODEL | No | Default model for the Gemini runner | |
| NEXUS_CLAUDE_MODELS | No | Comma-separated model list for the Claude runner | |
| NEXUS_GEMINI_MODELS | No | Comma-separated model list for the Gemini runner | |
| NEXUS_EXECUTION_MODE | No | Global execution mode (default or yolo) | default |
| NEXUS_OPENCODE_MODEL | No | Default model for the OpenCode runner | |
| NEXUS_OPENCODE_MODELS | No | Comma-separated model list for the OpenCode runner | |
| NEXUS_RETRY_MAX_DELAY | No | Maximum seconds to wait between retries | 60.0 |
| NEXUS_TIMEOUT_SECONDS | No | Subprocess timeout in seconds (10 minutes) | 600 |
| NEXUS_RETRY_BASE_DELAY | No | Base seconds for exponential backoff | 2.0 |
| NEXUS_OUTPUT_LIMIT_BYTES | No | Max output size in bytes before temp-file spillover | 50000 |
| NEXUS_RETRY_MAX_ATTEMPTS | No | Max attempts including the first (set to 1 to disable retries) | 3 |
| NEXUS_TOOL_TIMEOUT_SECONDS | No | Tool-level timeout in seconds (15 minutes); set to 0 to disable | 900 |
| NEXUS_CLI_DETECTION_TIMEOUT | No | Timeout in seconds for CLI binary version detection at startup | 30 |
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 | {
"listChanged": true
} |
| logging | {} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| promptB | Run one turn with a coding agent and return its final answer. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| backends_resource | Backend availability, models, and install hints. |
TDQS
Scored across 1 tool
With only a single tool, there is no possibility of confusing it with another, so misselection is impossible. However, there is no sibling tool to distinguish it from, making the dimension trivially satisfied rather than well-designed.
There is only one tool named 'prompt', a bare noun rather than a verb_noun pattern. With a single name there is nothing to be inconsistent with, but the naming itself is vague about what the tool does.
A single tool is far too thin for a server that wraps a coding agent. Core capabilities like starting a session, sending follow-ups, or checking status are absent, leaving the surface severely under-scoped.
Only one fire-and-forget turn is exposed, with no session lifecycle, multi-turn conversation, cancellation, or status operations. This leaves significant gaps for any realistic agent-driven coding workflow.