Skills | SignalWire

Skills are pluggable capabilities that add tools to your agent. Add a skill with add_skill() and it registers one or more SWAIG functions automatically. Skills handle setup, parameter validation, and tool registration so you can add features like weather, search, or math with a single call.

1 from signalwire import AgentBase
2 
3 class MyAgent(AgentBase):
4     def __init__(self):
5         super().__init__(name="my-agent")
6         self.set_prompt_text("You are a helpful assistant.")
7 
8         # No-config skill
9         self.add_skill("datetime")
10 
11         # Skill with parameters
12         self.add_skill("web_search", {
13             "api_key": "YOUR_KEY",
14             "search_engine_id": "YOUR_ENGINE_ID"
15         })
16 
17 if __name__ == "__main__":
18     MyAgent().run()

Skills Summary

Skill	Functions	API Required	Multi-Instance
`datetime`	2	No	No
`math`	1	No	No
`web_search`	1	Yes	Yes
`wikipedia_search`	1	No	No
`weather_api`	1	Yes	No
`joke`	1	Yes	No
`play_background_file`	1	No	Yes
`swml_transfer`	1	No	Yes
`datasphere`	1	Yes	Yes
`datasphere_serverless`	1	Yes	Yes
`native_vector_search`	1	No	Yes
`mcp_gateway`	Dynamic	No	Yes
`google_maps`	2	Yes	No
`info_gatherer`	2	No	Yes
`claude_skills`	Dynamic	No	Yes
`spider`	3	No	Yes
`api_ninjas_trivia`	1	Yes	Yes

Configuration

All skills accept configuration via a dictionary passed to add_skill(). Skills can also read values from environment variables when a parameter defines an env_var fallback.

1 # Direct configuration
2 self.add_skill("web_search", {"api_key": "KEY", "search_engine_id": "ID"})
3 
4 # Environment variable fallback
5 import os
6 self.add_skill("web_search", {
7     "api_key": os.getenv("GOOGLE_API_KEY"),
8     "search_engine_id": os.getenv("SEARCH_ENGINE_ID")
9 })

SWAIG Field Overrides

Override SWAIG function metadata for any skill by including a swaig_fields key:

1 from signalwire import AgentBase
2 
3 class MyAgent(AgentBase):
4     def __init__(self):
5         super().__init__(name="assistant", route="/assistant")
6         self.set_prompt_text("You are a helpful assistant.")
7         self.add_skill("datetime", {
8             "swaig_fields": {
9                 "fillers": {"en-US": ["Let me check the time...", "One moment..."]},
10                 "secure": False
11             }
12         })
13 
14 agent = MyAgent()
15 agent.serve()

Multi-Instance Skills

Skills that support multiple instances require unique tool_name values:

1 from signalwire import AgentBase
2 
3 class MyAgent(AgentBase):
4     def __init__(self):
5         super().__init__(name="assistant", route="/assistant")
6         self.set_prompt_text("You are a helpful assistant.")
7         self.add_skill("native_vector_search", {
8             "tool_name": "search_products",
9             "index_file": "/data/products.swsearch"
10         })
11         self.add_skill("native_vector_search", {
12             "tool_name": "search_faqs",
13             "index_file": "/data/faqs.swsearch"
14         })
15 
16 agent = MyAgent()
17 agent.serve()