agentica-sdk

Name: agentica-sdk
Author: parcadei

3.9kstars297forks

How to install agentica-sdk?

parcadei/continuous-claude-v3/agentica-sdk

$npx -y skills add parcadei/Continuous-Claude-v3 --skill agentica-sdk

Installs into the current project.

Summary

Build Python agents with Agentica SDK - @agentic decorator, spawn(), persistence, MCP integration

Files · 1

View on GitHub

SKILL.md

1	# Agentica SDK Reference (v0.3.1)
2
3	Build AI agents in Python using the Agentica framework. Agents can implement functions, maintain state, use tools, and coordinate with each other.
4
5	## When to Use
6
7	Use this skill when:
8	- Building new Python agents
9	- Adding agentic capabilities to existing code
10	- Integrating MCP tools with agents
11	- Implementing multi-agent orchestration
12	- Debugging agent behavior
13
14	## Quick Start
15
16	### Agentic Function (simplest)
17
18	```python
19	from agentica import agentic
20
21	@agentic()
22	async def add(a: int, b: int) -> int:
23	"""Returns the sum of a and b"""
24	...
25
26	result = await add(1, 2) # Agent computes: 3
27	```
28
29	### Spawned Agent (more control)
30
31	```python
32	from agentica import spawn
33
34	agent = await spawn(premise="You are a truth-teller.")
35	result: bool = await agent.call(bool, "The Earth is flat")
36	# Returns: False
37	```
38
39	## Core Patterns
40
41	### Return Types
42
43	```python
44	# String (default)
45	result = await agent.call("What is 2+2?")
46
47	# Typed output
48	result: int = await agent.call(int, "What is 2+2?")
49	result: dict[str, int] = await agent.call(dict[str, int], "Count items")
50
51	# Side-effects only
52	await agent.call(None, "Send message to John")
53	```
54
55	### Premise vs System Prompt
56
57	```python
58	# Premise: adds to default system prompt
59	agent = await spawn(premise="You are a math expert.")
60
61	# System: full control (replaces default)
62	agent = await spawn(system="You are a JSON-only responder.")
63	```
64
65	### Passing Tools (Scope)
66
67	```python
68	from agentica import agentic, spawn
69
70	# In decorator
71	@agentic(scope={'web_search': web_search_fn})
72	async def researcher(query: str) -> str:
73	"""Research a topic."""
74	...
75
76	# In spawn
77	agent = await spawn(
78	premise="Data analyzer",
79	scope={"analyze": custom_analyzer}
80	)
81
82	# Per-call scope
83	result = await agent.call(
84	dict[str, int],
85	"Analyze the dataset",
86	dataset=data, # Available as 'dataset'
87	analyzer=custom_fn # Available as 'analyzer'
88	)
89	```
90
91	### SDK Integration Pattern
92
93	```python
94	from slack_sdk import WebClient
95
96	slack = WebClient(token=SLACK_TOKEN)
97
98	# Extract specific methods
99	@agentic(scope={
100	'list_users': slack.users_list,
101	'send_message': slack.chat_postMessage
102	})
103	async def team_notifier(message: str) -> None:
104	"""Send team notifications."""
105	...
106	```
107
108	## Agent Instantiation
109
110	### spawn() - Async (most cases)
111
112	```python
113	agent = await spawn(premise="Helpful assistant")
114	```
115
116	### Agent() - Sync (for `__init__`)
117
118	```python
119	from agentica.agent import Agent
120
121	class CustomAgent:
122	def __init__(self):
123	# Synchronous - use Agent() not spawn()
124	self._brain = Agent(
125	premise="Specialized assistant",
126	scope={"tool": some_tool}
127	)
128
129	async def run(self, task: str) -> str:
130	return await self._brain(str, task)
131	```
132
133	## Model Selection
134
135	```python
136	# In spawn
137	agent = await spawn(
138	premise="Fast responses",
139	model="openai:gpt-5" # Default: openai:gpt-4.1
140	)
141
142	# In decorator
143	@agentic(model="anthropic:claude-sonnet-4.5")
144	async def analyze(text: str) -> dict:
145	"""Analyze text."""
146	...
147	```
148
149	Available models:
150	- `openai:gpt-3.5-turbo`, `openai:gpt-4o`, `openai:gpt-4.1`, `openai:gpt-5`
151	- `anthropic:claude-sonnet-4`, `anthropic:claude-opus-4.1`
152	- `anthropic:claude-sonnet-4.5`, `anthropic:claude-opus-4.5`
153	- Any OpenRouter slug (e.g., `google/gemini-2.5-flash`)
154
155	## Persistence (Stateful Agents)
156
157	```python
158	@agentic(persist=True)
159	async def chatbot(message: str) -> str:
160	"""Remembers conversation history."""
161	...
162
163	await chatbot("My name is Alice")
164	await chatbot("What's my name?") # Knows: Alice
165	```
166
167	For `spawn()` agents, state is automatic across calls to the same instance.
168
169	## Token Limits
170
171	```python
172	from agentica import spawn, MaxTokens
173
174	# Simple limit
175	agent = await spawn(
176	premise="Brief responses",
177	max_tokens=500
178	)
179
180	# Fine-grained control
181	agent = await spawn(
182	premise="Controlled output",
183	max_tokens=MaxTokens(
184	per_invocation=5000, # Total across all rounds
185	per_round=1000, # Per inference round
186	rounds=5 # Max inference rounds
187	)
188	)
189	```
190
191	## Token Usage Tracking
192
193	```python
194	from agentica import spawn, last_usage, total_usage
195
196	agent = await spawn(premise="You are helpful.")
197	await agent.call(str, "Hello!")
198
199	# Agent method
200	usage = agent.last_usage()
201	print(f"Last: {usage.input_tokens} in, {usage.output_tokens} out")
202
203	usage = agent.total_usage()
204	print(f"Total: {usage.total_tokens} processed")
205
206	# For @agentic functions
207	@agentic()
208	async def my_fn(x: str) -> str: ...
209
210	await my_fn("test")
211	print(last_usage(my_fn))
212	print(total_usage(my_fn))
213	```
214
215	## Streaming
216
217	```python
218	from agentica import spawn
219	from agentica.logging.loggers import StreamLogger
220	import asyncio
221
222	agent = await spawn(premise="You are helpful.")
223
224	stream = StreamLogger()
225	with stream:
226	result = asyncio.create_t

People who liked this also viewed

6 picks

Type

lark-openapi-explorer飞书/Lark 原生 OpenAPI 探索：从官方文档库中挖掘未经 CLI 封装的原生 OpenAPI 接口。当用户的需求无法被现有 lark-* skill 或 lark-cli 已注册命令满足，需要查找并调用原生飞书 OpenAPI 时使用。SkillsJul 2026378k16k
lark-skill-maker创建 lark-cli 的自定义 Skill。当用户需要把飞书 API 操作封装成可复用的 Skill（包装原子 API 或编排多步流程）时使用。SkillsJul 2026377k16k
azure-cost-optimizationOfficial agent plugin providing skills and MCP server configurations for Azure scenarios.SkillsJul 2026209k1.3k
vercel-react-native-skillsVercel's official collection of agent skillsSkillsJul 2026172k29k
mcp-builderGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate externaSkillsJul 202693k163k
gws-sharedgws CLI: Shared patterns for authentication, global flags, and output formatting.SkillsJul 202653k30k

agentica-sdk

3.9kstars297forks

View on GitHub

How to install agentica-sdk?

parcadei/continuous-claude-v3/agentica-sdk

$npx -y skills add parcadei/Continuous-Claude-v3 --skill agentica-sdk

Installs into the current project.

Summary

Build Python agents with Agentica SDK - @agentic decorator, spawn(), persistence, MCP integration

Files · 1

View on GitHub

SKILL.md

1	# Agentica SDK Reference (v0.3.1)
2
3	Build AI agents in Python using the Agentica framework. Agents can implement functions, maintain state, use tools, and coordinate with each other.
4
5	## When to Use
6
7	Use this skill when:
8	- Building new Python agents
9	- Adding agentic capabilities to existing code
10	- Integrating MCP tools with agents
11	- Implementing multi-agent orchestration
12	- Debugging agent behavior
13
14	## Quick Start
15
16	### Agentic Function (simplest)
17
18	```python
19	from agentica import agentic
20
21	@agentic()
22	async def add(a: int, b: int) -> int:
23	"""Returns the sum of a and b"""
24	...
25
26	result = await add(1, 2) # Agent computes: 3
27	```
28
29	### Spawned Agent (more control)
30
31	```python
32	from agentica import spawn
33
34	agent = await spawn(premise="You are a truth-teller.")
35	result: bool = await agent.call(bool, "The Earth is flat")
36	# Returns: False
37	```
38
39	## Core Patterns
40
41	### Return Types
42
43	```python
44	# String (default)
45	result = await agent.call("What is 2+2?")
46
47	# Typed output
48	result: int = await agent.call(int, "What is 2+2?")
49	result: dict[str, int] = await agent.call(dict[str, int], "Count items")
50
51	# Side-effects only
52	await agent.call(None, "Send message to John")
53	```
54
55	### Premise vs System Prompt
56
57	```python
58	# Premise: adds to default system prompt
59	agent = await spawn(premise="You are a math expert.")
60
61	# System: full control (replaces default)
62	agent = await spawn(system="You are a JSON-only responder.")
63	```
64
65	### Passing Tools (Scope)
66
67	```python
68	from agentica import agentic, spawn
69
70	# In decorator
71	@agentic(scope={'web_search': web_search_fn})
72	async def researcher(query: str) -> str:
73	"""Research a topic."""
74	...
75
76	# In spawn
77	agent = await spawn(
78	premise="Data analyzer",
79	scope={"analyze": custom_analyzer}
80	)
81
82	# Per-call scope
83	result = await agent.call(
84	dict[str, int],
85	"Analyze the dataset",
86	dataset=data, # Available as 'dataset'
87	analyzer=custom_fn # Available as 'analyzer'
88	)
89	```
90
91	### SDK Integration Pattern
92
93	```python
94	from slack_sdk import WebClient
95
96	slack = WebClient(token=SLACK_TOKEN)
97
98	# Extract specific methods
99	@agentic(scope={
100	'list_users': slack.users_list,
101	'send_message': slack.chat_postMessage
102	})
103	async def team_notifier(message: str) -> None:
104	"""Send team notifications."""
105	...
106	```
107
108	## Agent Instantiation
109
110	### spawn() - Async (most cases)
111
112	```python
113	agent = await spawn(premise="Helpful assistant")
114	```
115
116	### Agent() - Sync (for `__init__`)
117
118	```python
119	from agentica.agent import Agent
120
121	class CustomAgent:
122	def __init__(self):
123	# Synchronous - use Agent() not spawn()
124	self._brain = Agent(
125	premise="Specialized assistant",
126	scope={"tool": some_tool}
127	)
128
129	async def run(self, task: str) -> str:
130	return await self._brain(str, task)
131	```
132
133	## Model Selection
134
135	```python
136	# In spawn
137	agent = await spawn(
138	premise="Fast responses",
139	model="openai:gpt-5" # Default: openai:gpt-4.1
140	)
141
142	# In decorator
143	@agentic(model="anthropic:claude-sonnet-4.5")
144	async def analyze(text: str) -> dict:
145	"""Analyze text."""
146	...
147	```
148
149	Available models:
150	- `openai:gpt-3.5-turbo`, `openai:gpt-4o`, `openai:gpt-4.1`, `openai:gpt-5`
151	- `anthropic:claude-sonnet-4`, `anthropic:claude-opus-4.1`
152	- `anthropic:claude-sonnet-4.5`, `anthropic:claude-opus-4.5`
153	- Any OpenRouter slug (e.g., `google/gemini-2.5-flash`)
154
155	## Persistence (Stateful Agents)
156
157	```python
158	@agentic(persist=True)
159	async def chatbot(message: str) -> str:
160	"""Remembers conversation history."""
161	...
162
163	await chatbot("My name is Alice")
164	await chatbot("What's my name?") # Knows: Alice
165	```
166
167	For `spawn()` agents, state is automatic across calls to the same instance.
168
169	## Token Limits
170
171	```python
172	from agentica import spawn, MaxTokens
173
174	# Simple limit
175	agent = await spawn(
176	premise="Brief responses",
177	max_tokens=500
178	)
179
180	# Fine-grained control
181	agent = await spawn(
182	premise="Controlled output",
183	max_tokens=MaxTokens(
184	per_invocation=5000, # Total across all rounds
185	per_round=1000, # Per inference round
186	rounds=5 # Max inference rounds
187	)
188	)
189	```
190
191	## Token Usage Tracking
192
193	```python
194	from agentica import spawn, last_usage, total_usage
195
196	agent = await spawn(premise="You are helpful.")
197	await agent.call(str, "Hello!")
198
199	# Agent method
200	usage = agent.last_usage()
201	print(f"Last: {usage.input_tokens} in, {usage.output_tokens} out")
202
203	usage = agent.total_usage()
204	print(f"Total: {usage.total_tokens} processed")
205
206	# For @agentic functions
207	@agentic()
208	async def my_fn(x: str) -> str: ...
209
210	await my_fn("test")
211	print(last_usage(my_fn))
212	print(total_usage(my_fn))
213	```
214
215	## Streaming
216
217	```python
218	from agentica import spawn
219	from agentica.logging.loggers import StreamLogger
220	import asyncio
221
222	agent = await spawn(premise="You are helpful.")
223
224	stream = StreamLogger()
225	with stream:
226	result = asyncio.create_t

People who liked this also viewed

6 picks

Type