codebolt.agent
Agent management and lifecycle operations for controlling and coordinating AI agents.
Available Tools
agent_start
- Start a new agent instanceagent_find
- Find existing agents by criteriaagent_list
- List all active agentsagent_stop
- Stop a running agentagent_install
- Install a new agentattempt_completion
- Mark agent task as completed
Sample Usage
// Start a new agent
const startResult = await codeboltMCP.executeTool(
"codebolt.agent",
"agent_start",
{
agentType: "codeAnalyzer",
config: { language: "javascript" }
}
);
// List all active agents
const listResult = await codeboltMCP.executeTool(
"codebolt.agent",
"agent_list",
{}
);
// Find specific agents
const findResult = await codeboltMCP.executeTool(
"codebolt.agent",
"agent_find",
{
criteria: { type: "codeAnalyzer", status: "active" }
}
);
// Stop an agent
const stopResult = await codeboltMCP.executeTool(
"codebolt.agent",
"agent_stop",
{ agentId: "agent-123" }
);
// Install a new agent
const installResult = await codeboltMCP.executeTool(
"codebolt.agent",
"agent_install",
{
source: "marketplace",
agentName: "testRunner"
}
);
// Mark task completion
const completeResult = await codeboltMCP.executeTool(
"codebolt.agent",
"attempt_completion",
{
agentId: "agent-123",
result: "Task completed successfully"
}
);
info
This functionality provides comprehensive agent lifecycle management for orchestrating multiple AI agents.