scaffolding-openai-agents

Name: scaffolding-openai-agents
Author: mjunaidca

20installs11forks

How to install scaffolding-openai-agents?

mjunaidca/mjs-agent-skills/scaffolding-openai-agents

$npx -y skills add mjunaidca/mjs-agent-skills --skill scaffolding-openai-agents

Installs into the current project.

Summary

Builds AI agents using OpenAI Agents SDK with async/await patterns and multi-agent orchestration. Use when creating tutoring agents, building agent handoffs, implementing tool-calling agents, or orchestrating multiple specialists. Covers Agent class, Runner patterns, function too

Files · 1

View on GitHub

SKILL.md

1	# Scaffolding OpenAI Agents
2
3	Build production AI agents using OpenAI Agents SDK with native async/await patterns.
4
5	## Quick Start
6
7	```bash
8	# Project setup
9	mkdir my-agent && cd my-agent
10	python -m venv .venv && source .venv/bin/activate
11	pip install openai-agents
12
13	# Set API key
14	export OPENAI_API_KEY=sk-...
15	```
16
17	```python
18	# main.py
19	import asyncio
20	from agents import Agent, Runner
21
22	agent = Agent(
23	name="Python Tutor",
24	instructions="You help students learn Python. Explain concepts clearly with examples."
25	)
26
27	async def main():
28	result = await Runner.run(agent, "Explain list comprehensions")
29	print(result.final_output)
30
31	asyncio.run(main())
32	```
33
34	## Agent Configuration
35
36	### Basic Agent
37
38	```python
39	from agents import Agent
40
41	tutor = Agent(
42	name="Python Tutor",
43	instructions="""You are an expert Python tutor.
44	Explain concepts clearly with examples.
45	Ask clarifying questions when needed.
46	Provide practice exercises after explanations.""",
47	model="gpt-4o"
48	)
49	```
50
51	### With Model Settings
52
53	```python
54	from agents import Agent, ModelSettings
55
56	agent = Agent(
57	name="Creative Writer",
58	instructions="Write creative stories based on prompts.",
59	model="gpt-4o",
60	model_settings=ModelSettings(
61	temperature=0.9,
62	max_tokens=2000
63	)
64	)
65	```
66
67	### With Structured Output
68
69	```python
70	from pydantic import BaseModel
71	from agents import Agent
72
73	class CodeReview(BaseModel):
74	issues: list[str]
75	suggestions: list[str]
76	score: int
77
78	reviewer = Agent(
79	name="Code Reviewer",
80	instructions="Review Python code for issues and improvements.",
81	output_type=CodeReview # Forces structured JSON output
82	)
83	```
84
85	## Runner Patterns
86
87	### Async Run (Primary)
88
89	```python
90	import asyncio
91	from agents import Agent, Runner
92
93	async def main():
94	agent = Agent(name="Helper", instructions="Be helpful")
95
96	# Single query
97	result = await Runner.run(agent, "What is Python?")
98	print(result.final_output)
99
100	# With conversation history
101	messages = [
102	{"role": "user", "content": "My name is Alex"},
103	{"role": "assistant", "content": "Nice to meet you, Alex!"},
104	{"role": "user", "content": "What's my name?"}
105	]
106	result = await Runner.run(agent, messages)
107	print(result.final_output) # "Your name is Alex"
108
109	asyncio.run(main())
110	```
111
112	### Sync Run (Simple Scripts)
113
114	```python
115	from agents import Agent, Runner
116
117	agent = Agent(name="Helper", instructions="Be helpful")
118	result = Runner.run_sync(agent, "Hello!")
119	print(result.final_output)
120	```
121
122	### Streaming Run
123
124	```python
125	import asyncio
126	from agents import Agent, Runner
127
128	async def main():
129	agent = Agent(name="Storyteller", instructions="Tell engaging stories")
130
131	result = Runner.run_streamed(agent, "Tell me a short story")
132
133	async for event in result.stream_events():
134	if hasattr(event, 'delta'):
135	print(event.delta, end='', flush=True)
136
137	print() # Newline at end
138
139	asyncio.run(main())
140	```
141
142	### Conversation Continuation
143
144	```python
145	async def chat_session():
146	agent = Agent(name="Tutor", instructions="You are a Python tutor")
147
148	# First turn
149	result1 = await Runner.run(agent, "Explain decorators")
150	print(f"Tutor: {result1.final_output}")
151
152	# Continue conversation
153	messages = result1.to_input_list() + [
154	{"role": "user", "content": "Show me an example"}
155	]
156	result2 = await Runner.run(agent, messages)
157	print(f"Tutor: {result2.final_output}")
158	```
159
160	## Function Tools
161
162	### Basic Tool
163
164	```python
165	from agents import Agent, function_tool
166
167	@function_tool
168	def get_current_time() -> str:
169	"""Get the current time."""
170	from datetime import datetime
171	return datetime.now().strftime("%H:%M:%S")
172
173	@function_tool
174	def calculate(expression: str) -> float:
175	"""Calculate a mathematical expression.
176
177	Args:
178	expression: A valid Python math expression like "2 + 2" or "10 * 5"
179	"""
180	return eval(expression) # Use safe_eval in production
181
182	agent = Agent(
183	name="Assistant",
184	instructions="Help with calculations and time queries.",
185	tools=[get_current_time, calculate]
186	)
187	```
188
189	### Async Tool
190
191	```python
192	import httpx
193	from agents import Agent, function_tool
194
195	@function_tool
196	async def fetch_weather(city: str) -> str:
197	"""Fetch current weather for a city.
198
199	Args:
200	city: The city name to get weather for
201	"""
202	async with httpx.AsyncClient() as client:
203	response = await client.get(
204	f"https://wttr.in/{city}?format=3"
205	)
206	return response.text
207
208	agent = Agent(
209	name="Weather B

People who liked this also viewed

6 picks

Type

scaffolding-openai-agents

20installs11forks

View on GitHub

How to install scaffolding-openai-agents?

mjunaidca/mjs-agent-skills/scaffolding-openai-agents

$npx -y skills add mjunaidca/mjs-agent-skills --skill scaffolding-openai-agents

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Scaffolding OpenAI Agents
2
3	Build production AI agents using OpenAI Agents SDK with native async/await patterns.
4
5	## Quick Start
6
7	```bash
8	# Project setup
9	mkdir my-agent && cd my-agent
10	python -m venv .venv && source .venv/bin/activate
11	pip install openai-agents
12
13	# Set API key
14	export OPENAI_API_KEY=sk-...
15	```
16
17	```python
18	# main.py
19	import asyncio
20	from agents import Agent, Runner
21
22	agent = Agent(
23	name="Python Tutor",
24	instructions="You help students learn Python. Explain concepts clearly with examples."
25	)
26
27	async def main():
28	result = await Runner.run(agent, "Explain list comprehensions")
29	print(result.final_output)
30
31	asyncio.run(main())
32	```
33
34	## Agent Configuration
35
36	### Basic Agent
37
38	```python
39	from agents import Agent
40
41	tutor = Agent(
42	name="Python Tutor",
43	instructions="""You are an expert Python tutor.
44	Explain concepts clearly with examples.
45	Ask clarifying questions when needed.
46	Provide practice exercises after explanations.""",
47	model="gpt-4o"
48	)
49	```
50
51	### With Model Settings
52
53	```python
54	from agents import Agent, ModelSettings
55
56	agent = Agent(
57	name="Creative Writer",
58	instructions="Write creative stories based on prompts.",
59	model="gpt-4o",
60	model_settings=ModelSettings(
61	temperature=0.9,
62	max_tokens=2000
63	)
64	)
65	```
66
67	### With Structured Output
68
69	```python
70	from pydantic import BaseModel
71	from agents import Agent
72
73	class CodeReview(BaseModel):
74	issues: list[str]
75	suggestions: list[str]
76	score: int
77
78	reviewer = Agent(
79	name="Code Reviewer",
80	instructions="Review Python code for issues and improvements.",
81	output_type=CodeReview # Forces structured JSON output
82	)
83	```
84
85	## Runner Patterns
86
87	### Async Run (Primary)
88
89	```python
90	import asyncio
91	from agents import Agent, Runner
92
93	async def main():
94	agent = Agent(name="Helper", instructions="Be helpful")
95
96	# Single query
97	result = await Runner.run(agent, "What is Python?")
98	print(result.final_output)
99
100	# With conversation history
101	messages = [
102	{"role": "user", "content": "My name is Alex"},
103	{"role": "assistant", "content": "Nice to meet you, Alex!"},
104	{"role": "user", "content": "What's my name?"}
105	]
106	result = await Runner.run(agent, messages)
107	print(result.final_output) # "Your name is Alex"
108
109	asyncio.run(main())
110	```
111
112	### Sync Run (Simple Scripts)
113
114	```python
115	from agents import Agent, Runner
116
117	agent = Agent(name="Helper", instructions="Be helpful")
118	result = Runner.run_sync(agent, "Hello!")
119	print(result.final_output)
120	```
121
122	### Streaming Run
123
124	```python
125	import asyncio
126	from agents import Agent, Runner
127
128	async def main():
129	agent = Agent(name="Storyteller", instructions="Tell engaging stories")
130
131	result = Runner.run_streamed(agent, "Tell me a short story")
132
133	async for event in result.stream_events():
134	if hasattr(event, 'delta'):
135	print(event.delta, end='', flush=True)
136
137	print() # Newline at end
138
139	asyncio.run(main())
140	```
141
142	### Conversation Continuation
143
144	```python
145	async def chat_session():
146	agent = Agent(name="Tutor", instructions="You are a Python tutor")
147
148	# First turn
149	result1 = await Runner.run(agent, "Explain decorators")
150	print(f"Tutor: {result1.final_output}")
151
152	# Continue conversation
153	messages = result1.to_input_list() + [
154	{"role": "user", "content": "Show me an example"}
155	]
156	result2 = await Runner.run(agent, messages)
157	print(f"Tutor: {result2.final_output}")
158	```
159
160	## Function Tools
161
162	### Basic Tool
163
164	```python
165	from agents import Agent, function_tool
166
167	@function_tool
168	def get_current_time() -> str:
169	"""Get the current time."""
170	from datetime import datetime
171	return datetime.now().strftime("%H:%M:%S")
172
173	@function_tool
174	def calculate(expression: str) -> float:
175	"""Calculate a mathematical expression.
176
177	Args:
178	expression: A valid Python math expression like "2 + 2" or "10 * 5"
179	"""
180	return eval(expression) # Use safe_eval in production
181
182	agent = Agent(
183	name="Assistant",
184	instructions="Help with calculations and time queries.",
185	tools=[get_current_time, calculate]
186	)
187	```
188
189	### Async Tool
190
191	```python
192	import httpx
193	from agents import Agent, function_tool
194
195	@function_tool
196	async def fetch_weather(city: str) -> str:
197	"""Fetch current weather for a city.
198
199	Args:
200	city: The city name to get weather for
201	"""
202	async with httpx.AsyncClient() as client:
203	response = await client.get(
204	f"https://wttr.in/{city}?format=3"
205	)
206	return response.text
207
208	agent = Agent(
209	name="Weather B

People who liked this also viewed

6 picks

Type