菜单
Teaches how to coordinate multiple specialized agents, distribute tasks, establish consensus, and manage role-based multi-agent swarms using LollmsSwarm.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Teaches how to build, discover, and execute local, zero-configuration Python tools in Lollms using AST schema parsing and context injection. Includes step-by-step tool file construction guide and code examples.
Teaches the model about the persistent tiered memory system (Level 1 Working, Level 2 Deep, Level 3 Archive), proactive keyword pulling, and the required tagging protocols.
Teaches the model how to formulate strategic <coding_plan> blocks, apply indentation-agnostic Aider search/replace patches, manage Git-like version logs, and control in-message status reporting.
Teaches the model how to manage stateful discussions, track conversational trees/branches, navigate alternative sibling replies, and execute context-aware generation.
Teaches how to import, parse, and process PDF, DOCX, PPTX, and image files using anchor-mapped and OCR import modes in LollmsDiscussion.
Guides the model on performing basic text generation, counting tokens, detokenizing lists, and generating structurally validated JSON outputs with LollmsClient.
| name | Lollms Swarms and Multi-Agent Orchestration |
| description | Teaches how to coordinate multiple specialized agents, distribute tasks, establish consensus, and manage role-based multi-agent swarms using LollmsSwarm. |
| author | ParisNeo |
| version | 1.0.0 |
| category | lollms_client/swarm |
| created | "2026-05-24T00:00:00.000Z" |
This skill explains how to build, configure, and orchestrate role-based multi-agent swarms using the lollms_client library.
An Agent is a stateful entity assigned a specific AgentRole (e.g. DOMAIN_EXPERT, AUDITOR, ORCHESTRATOR) and a custom personality/system prompt.
Instantiate an agent using a loaded LollmsClient and custom metadata.
from lollms_client import LollmsClient
from lollms_client.lollms_agent import Agent, AgentRole
from lollms_client.lollms_personality import LollmsPersonality
client = LollmsClient(llm_binding_name="ollama", llm_binding_config={"model_name": "gemma4:e2b"})
# Define expert personality
python_expert = LollmsPersonality(
name="PythonExpert",
system_prompt="You are a Senior Python Architect. Focus on clean code, SOLID principles, and type-safety."
)
# Instantiate the Agent
agent = Agent(
lc=client,
personality=python_expert,
name="PythonExpert",
role=AgentRole.DOMAIN_EXPERT,
model_params={"temperature": 0.2}
)
The SwarmOrchestrator manages collaborative task distribution across several specialized agents. It handles the round-robin discussions, collects intermediate answers, and facilitates the consensus pass.
from lollms_client.lollms_swarm import SwarmOrchestrator, SwarmConfig
# Configure the swarm
swarm_config = SwarmConfig(
max_rounds=3,
consensus_threshold=0.85,
enable_peer_review=True
)
# Initialize the orchestrator inside a discussion
orchestrator = SwarmOrchestrator(
discussion=discussion,
agents=[agent_developer, agent_auditor],
config=swarm_config
)
# Dispatch a query to the swarm
result = orchestrator.run("Design a thread-safe connection pool for SQLite.")
print(f"Final Answer: {result['response']}")
During execution, the swarm communicates progress through specific message types:
MSG_TYPE_SWARM_AGENT_START: An agent begins generating its turn.MSG_TYPE_SWARM_AGENT_END: An agent finishes its generation.MSG_TYPE_SWARM_CONSENSUS: The orchestrator evaluates the similarity of responses and reports on consensus strength.