- name
SKILL.md
技能文件保留在发布者的仓库中。
在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add coco-research/coco --skill agent-lightning -a cursor -y在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add coco-research/coco --skill agent-lightning -a claude-code -y在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add coco-research/coco --skill agent-lightning -a codex -y在终端中粘贴复制的命令。它会把这个 SKILL.md 安装到对应的智能体中。
npx skills add coco-research/coco --skill agent-lightning -a grok -y在 Grok Bot 聊天中发送复制的消息。Bot 会从 GitHub 获取这个 SKILL.md,并添加到你所有 Bot 共享的技能中。
Install this skill: https://github.com/coco-research/coco/tree/HEAD/skills/agent-lightning来自发布者
Microsoft's framework for training AI agents with reinforcement learning, automatic prompt optimization, and supervised fine-tuning.
pip install agentlightning
For nightly builds:
pip install --upgrade --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ --pre agentlightning
Add agl.emit_xxx() helpers to your existing agent:
import agentlightning as agl
# Your existing agent code
def my_agent(task):
agl.emit_input(task) # Track input
response = llm.generate(task)
agl.emit_output(response) # Track output
reward = evaluate(response)
agl.emit_reward(reward) # Track reward
return response
Agent (your code) → agl.emit_xxx() → Spans → LightningStore → Algorithm → Updated Resources
import agentlightning as agl
# Basic emissions
agl.emit_input(prompt) # Track input to agent
agl.emit_output(response) # Track agent output
agl.emit_reward(score) # Track reward signal
agl.emit_tool_call(name, args) # Track tool usage
agl.emit_tool_result(result) # Track tool results
from agentlightning import Tracer
tracer = Tracer(store=store)
with tracer.trace_context(task_id="task-123"):
# All emissions within this context are grouped
result = agent.run(task)
# Retrieve trace after execution
trace = tracer.get_last_trace()
Agent Lightning integrates with OpenTelemetry:
from agentlightning.utils.otel import get_tracer
tracer = get_tracer() # Returns OTel tracer for "agentlightning"
from agentlightning.store.memory import InMemoryLightningStore
store = InMemoryLightningStore()
from agentlightning.store.client_server import (
LightningStoreServer,
LightningStoreClient
)
# Server side
server = LightningStoreServer(store, host="0.0.0.0", port=8080)
await server.start()
# Client side
client = LightningStoreClient("http://localhost:8080")
# Add rollouts (tasks for the agent)
await store.enqueue_rollout(task=task, config=RolloutConfig())
# Query rollouts
rollouts = await store.query_rollouts(status_in=["completed"])
# Add resources (updated prompts, weights)
await store.add_resources(resources)
# Get latest resources
resources = await store.get_latest_resources()
import agentlightning as agl
trainer = agl.Trainer(
n_runners=8, # Parallel rollout workers
algorithm=algorithm, # Your chosen algorithm
store=store # Optional, creates InMemory if not provided
)
trainer.run()
from agentlightning import LightningStore
from agentlightning.types import ExecutionEvent
async def my_algorithm(store: LightningStore, event: ExecutionEvent):
# Fetch completed rollouts
rollouts = await store.query_rollouts(status_in=["completed"])
# Process traces, compute gradients, etc.
new_resources = optimize(rollouts)
# Push updated resources
await store.add_resources(new_resources)
async def my_runner(store: LightningStore, worker_id: int, event: ExecutionEvent):
while not event.is_set():
rollout = await store.dequeue_rollout()
if rollout:
result = execute_task(rollout.task)
await store.update_rollout(
rollout_id=rollout.id,
status="completed",
result=result
)
For RL training with vLLM backend:
from agentlightning.algorithm.verl import VeRLAlgorithm
algorithm = VeRLAlgorithm(
model="your-model",
learning_rate=1e-5,
batch_size=32
)
from agentlightning.algorithm.apo import APOAlgorithm
algorithm = APOAlgorithm(
optimizer_model="gpt-4",
target_model="gpt-3.5-turbo"
)
from agentlightning.instrumentation.langchain import instrument_langchain
instrument_langchain() # Auto-traces all LangChain calls
from agentlightning.instrumentation.openai import instrument_openai
instrument_openai() # Auto-traces OpenAI API calls
from agentlightning.instrumentation.vllm import instrument_vllm
instrument_vllm() # Instrument vLLM for token-level tracing
from agentlightning import setup_logging
setup_logging(
level="DEBUG",
submodule_levels={
"agentlightning.store": "INFO",
"agentlightning.tracer": "DEBUG"
}
)
Agent Lightning emits Prometheus-compatible metrics:
agl.store.total - Store operation countsagl.store.latency - Store operation latenciesagl.rollouts.total - Rollout counts by statusagl.rollouts.duration - Rollout execution timesdef compute_reward(task, response):
"""Good rewards are: normalized, dense when possible, aligned with goals."""
correctness = check_correctness(task, response) # 0-1
efficiency = measure_efficiency(response) # 0-1
return 0.7 * correctness + 0.3 * efficiency
Train specific agents in a multi-agent system:
with tracer.trace_context(agent_id="planner"):
plan = planner.run(task)
with tracer.trace_context(agent_id="executor"):
result = executor.run(plan)
# Only the executor's traces are used for training
# Save checkpoint
await store.add_resources(
checkpoint=True,
resources=current_resources
)
# Load latest
resources = await store.get_latest_resources()
For JavaScript/TypeScript agents (like Claude-based apps), you have two options:
Create a Python microservice that:
Use LightningStoreServer as a REST backend:
// JavaScript client
const response = await fetch('http://localhost:8080/rollouts', {
method: 'POST',
body: JSON.stringify({
task: { prompt: userMessage },
config: { max_retries: 3 }
})
});
Train and optimize AI agents using Microsoft's Agent Lightning framework with reinforcement learning. Use when setting up agent training, instrumenting agents with tracing, configuring LightningStore,
---name: agent-lightningdescription: Train and optimize AI agents using Microsoft's Agent Lightning framework with reinforcement learning. Use when…---# skills/agent-lightning/SKILL.md概览
Train and optimize AI agents using Microsoft's Agent Lightning framework with reinforcement learning. Use when setting up agent training, instrumenting agents with tracing, configuring LightningStore, implementing reward functions, or optimizing prompts with RL/APO algorithms.
常见问题
从发布者处下载 skills/agent-lightning/SKILL.md,或复制路径后打开仓库。PlugMCP 只收录该技能,不托管文件。
发布者将它存放在 https://github.com/coco-research/coco。
Claude、Cursor、Codex 和 Grok Build 可以读取你从发布者仓库添加的 SKILL.md。在聊天中发送技能的 GitHub 文件夹链接,Grok Bot 就会安装它。