polyagent
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| GOOGLE_API_KEY | Yes | API key for Google Gemini. | |
| OLLAMA_API_KEY | Yes | API key for Ollama Cloud. | |
| OLLAMA_BASE_URL | Yes | Base URL for Ollama Cloud. | https://ollama.com/v1 |
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 |
|---|---|
| register_agentA | Register a new AI agent. Args: name: Unique name for the agent provider: Provider type (openai_compat, anthropic, bedrock, gemini) model: Model identifier (e.g., 'glm-4', 'claude-3-5-sonnet') system_prompt: System prompt for the agent description: Human-readable description api_key_env: Environment variable name for API key temperature: Sampling temperature (0.0-2.0) max_tokens: Maximum tokens to generate capabilities: List of agent capabilities base_url: Base URL for OpenAI-compatible providers region: AWS region for Bedrock Returns: Registered agent profile |
| list_agentsB | List all registered agents. Args: capability: Optional capability filter Returns: List of agent profiles |
| update_agentC | Update an existing agent. Args: name: Agent name **kwargs: Fields to update Returns: Updated agent profile |
| remove_agentC | Remove an agent. Args: name: Agent name Returns: Removed agent profile |
| execute_agentC | Execute a single agent. Args: agent_name: Name of the agent to execute input_content: Input content to process context: Optional context from Claude Returns: Execution result |
| execute_pipelineB | Execute a pipeline with multiple agents. Args:
agents: List of agent names
input_content: Input content to process
mode: Pipeline mode (sequential, iterative, parallel)
context: Optional context from Claude
max_iterations: Maximum iterations for iterative mode
confidence_threshold: Confidence threshold for stopping
require_approval_after: Pause iterative mode for human approval after N
iterations. Defaults to None (disabled) so Returns: Pipeline execution results |
| set_safety_configA | Configure loop prevention and safety settings. Args: max_iterations: Maximum iterations before stopping confidence_threshold: Stop when confidence exceeds threshold require_approval_after: Require approval after N iterations timeout_seconds: Global timeout in seconds Returns: Updated safety configuration |
| get_safety_configA | Get current safety configuration. Returns: Current safety configuration |
| get_available_modelsA | Get list of all available models and their capabilities. This tool should be called by Claude to discover which models are available and what tasks they're best suited for. Returns: Dictionary with model registry including capabilities |
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 9 tools
Every tool targets a distinct concern: agent lifecycle, agent execution, pipeline execution, safety configuration, and model discovery. The only potentially similar pair (execute_agent vs execute_pipeline) is clearly differentiated by scope: single agent vs multi-agent pipeline.
Tool names follow a consistent verb_noun pattern: list_, register_, update_, remove_, execute_, get_, set_. Naming clearly indicates both the action and the resource, with no mixed conventions or vague verbs.
Nine tools is well-scoped for an agent management and orchestration server. Each tool covers a necessary aspect of the domain without redundancy or bloat.
The tool surface provides complete agent lifecycle coverage (list, register, update, remove), execution paths for both individual and multi-agent workflows, safety controls, and model discovery. There are no obvious dead ends or missing operations for the stated purpose.