Lad MCP Server
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| LAD_ENV_FILE | No | Optional env file path | |
| DEEPSEEK_API_KEY | No | Optional API key for DeepSeek | |
| KIMI_CODE_API_KEY | No | Optional API key for Kimi Code | |
| OPENROUTER_API_KEY | Yes | API key for OpenRouter | |
| OPENROUTER_X_TITLE | No | Optional X-Title forwarded to OpenRouter | |
| ZAI_CODING_PLAN_KEY | No | Optional API key for Z.AI Coding Plan | |
| OPENROUTER_HTTP_REFERER | No | Optional HTTP referer forwarded to OpenRouter | |
| INTERMITTENT_REVIEW_CALLS | No | Number of intermittent review calls | 2 |
| LAD_SERENA_MAX_TOOL_CALLS | No | Maximum Serena tool calls | 32 |
| LAD_SERENA_MAX_DIR_ENTRIES | No | Maximum directory entries | 100 |
| LAD_SERENA_MAX_TOTAL_CHARS | No | Maximum total characters | 50000 |
| OPENROUTER_MAX_INPUT_CHARS | No | Maximum input characters | 100000 |
| OPENROUTER_INCLUDE_REASONING | No | Whether to include reasoning in output | false |
| LAD_SERENA_MAX_SEARCH_RESULTS | No | Maximum search results | 20 |
| OPENROUTER_FIXED_OUTPUT_TOKENS | No | Fixed output tokens for OpenRouter | 8192 |
| LAD_SERENA_TOOL_TIMEOUT_SECONDS | No | Timeout for Serena tool calls | 30 |
| LAD_SERENA_MAX_TOOL_RESULT_CHARS | No | Maximum tool result characters | 12000 |
| OPENROUTER_PRIMARY_REVIEWER_MODEL | No | Primary reviewer model | moonshotai/kimi-k2.5 |
| OPENROUTER_CONTEXT_OVERHEAD_TOKENS | No | Context overhead tokens for OpenRouter | 2000 |
| OPENROUTER_MAX_CONCURRENT_REQUESTS | No | Maximum concurrent requests to OpenRouter | 4 |
| OPENROUTER_REVIEWER_TIMEOUT_SECONDS | No | Wall-clock timeout per reviewer run | 300 |
| OPENROUTER_SECONDARY_REVIEWER_MODEL | No | Secondary reviewer model, set to 0 to disable | minimax/minimax-m2.7 |
| OPENROUTER_TOOL_CALL_TIMEOUT_SECONDS | No | Timeout per tool call | 360 |
| OPENROUTER_MODEL_METADATA_TTL_SECONDS | No | TTL for model metadata caching | 3600 |
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": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| system_design_reviewA | Review a system design proposal and constraints using two LLM reviewers in parallel. At least one of |
| code_reviewA | Review a code snippet or diff using two LLM reviewers in parallel. At least one of |
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 2 tools
The two tools, code_review and system_design_review, target distinctly different domains: code snippets/diffs vs. system design proposals. There is no overlap in their purpose, making disambiguation straightforward.
Both tools follow a consistent verb_noun pattern using snake_case (code_review, system_design_review). The naming is predictable and clear.
With only two tools, the server is quite minimal. While the tools are specific and well-defined, the narrow scope may feel thin for a general development assistant, but it is not excessively small for a focused purpose.
The server covers two common review types (code and system design) but lacks additional reviews like architecture, security, or documentation. The tools accept file paths, suggesting some flexibility, but the surface is notably incomplete for comprehensive review workflows.