leam-opt-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| OPENAI_MODEL | No | OpenAI vision model (required if ANTENNA_VISION_PROVIDER is 'openai'). | |
| DEEPSEEK_MODEL | No | DeepSeek model for text processing. | |
| OPENAI_API_KEY | No | API key for OpenAI (required if ANTENNA_VISION_PROVIDER is 'openai'). | |
| OLLAMA_BASE_URL | No | Base URL for Ollama server. | http://localhost:11434 |
| DEEPSEEK_API_KEY | Yes | API key for DeepSeek. | |
| DEEPSEEK_BASE_URL | No | Base URL for DeepSeek API. | https://api.deepseek.com |
| OLLAMA_VISION_MODEL | No | Ollama vision model, e.g., 'qwen3-vl:8b'. | |
| ANTENNA_MCP_WORKSPACE | No | Directory for MCP workspace files. | .antenna-mcp |
| ANTENNA_TEXT_PROVIDER | Yes | Text provider to use, e.g., 'deepseek'. | |
| ANTENNA_VISION_PROVIDER | Yes | Vision provider to use: 'ollama' or 'openai'. | |
| ANTENNA_MCP_ALLOW_SIMULATION | No | Set to '1' to allow HFSS simulation execution. | 0 |
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 |
|---|---|
| antenna_server_healthA | Report configured workspace, optional backend availability, and execution gate. |
| create_antenna_modeling_jobB | Create a staged prompt-driven modeling job without running the LLM or simulator. |
| run_antenna_modeling_jobB | Run modeling stages through the requested stage and save all intermediate artifacts. |
| generate_antenna_pythonA | Generate a complete Python model artifact without starting AEDT or using a license. The returned file is safe to import offline. It exposes |
| submit_antenna_model_feedbackA | Record the user's HFSS/image comparison notes without running or changing AEDT. |
| regenerate_antenna_python_from_feedbackB | Use recorded feedback to produce the next versioned Python model, still without AEDT. |
| analyze_antenna_sourceC | Recognize topology, labels, dimensions, materials, and uncertainties in antenna images/PDFs. |
| refine_antenna_sourceB | Reconcile raw visual evidence with extracted text; produces a hash-frozen review candidate. |
| approve_antenna_sourceB | Approve the reviewed source candidate before downstream model/code generation. |
| recheck_antenna_sourceB | Apply a reviewed source audit deterministically and issue a new hash-frozen candidate. |
| propose_antenna_engineering_assumptionB | Propose a value for a null/unresolved source parameter and return its review hash. |
| approve_antenna_engineering_assumptionB | Approve the exact hash-frozen engineering assumption candidate after user review. |
| compile_reviewed_antenna_modelC | Deterministically compile approved evidence and assumptions into reviewed HFSS artifacts. |
| prepare_antenna_artifact_reviewA | Hash every generated artifact and return the approval token required for HFSS execution. |
| build_hfss_projectC | Build reviewed artifacts; any edit after review invalidates the supplied approval hash. |
| create_hfss_optimization_jobB | Copy an existing HFSS project into an isolated job and prepare black-box optimization. |
| run_hfss_optimization_jobB | Run HFSS trials. Requires ANTENNA_MCP_ALLOW_SIMULATION=1 and never overwrites the source project. |
| get_antenna_jobB | Read current job state and artifact paths. |
| create_antenna_pipelineC | Create one end-to-end job spanning multimodal understanding, HFSS build, and optimization. |
| generate_antenna_pipelineC | Run source understanding through optimization planning, then stop at the artifact review gate. |
| build_approved_antenna_pipelineC | Build the approved pipeline model and stop before expensive optimization. |
| optimize_antenna_pipelineC | Run the planned HFSS optimization and produce the best project plus complete trial history. |
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 22 tools
Multiple tools involve approval hashes and source refinement, making boundaries unclear (e.g., approve_antenna_source vs approve_antenna_engineering_assumption vs prepare_antenna_artifact_review). Pairs like refine_antenna_source and recheck_antenna_source appear to do nearly the same thing, causing likely misselection.
Naming mixes antenna and hfss prefixes (create_antenna_modeling_job vs build_hfss_project) and includes a noun-phrase tool (antenna_server_health). Verb forms vary (create, generate, build, compile) and long names like regenerate_antenna_python_from_feedback break any clean pattern.
With 22 tools, the server is on the heavy side, though the complexity of the domain (source analysis, modeling, simulation, optimization) justifies many steps. However, the number feels bloated due to overlapping tools that could be consolidated.
The toolset covers a full pipeline from source analysis through approval, Python generation, feedback, and HFSS optimization. Minor gaps exist (e.g., no job listing/cancellation, no pipeline status), but the core workflow is thoroughly covered.