yr.tool
tool(
fn: Callable | None = None,
creds: str | list[str] | None = None,
insert_result: bool = True,
) → DecoratedTool | Callable[[Callable], DecoratedTool]Decorate a Python function as a Yera tool.
Tools are self-contained chunks of Python an
@app can call into — the app's
LLM drives control flow, deciding when and with what arguments to
invoke each tool. Usable bare (@tool) or with keyword arguments
(@tool(creds=...)).
If creds is provided, the named credential groups are resolved
from the active Yera profile and bound to the tool's context
before the function body runs; the body reads them back via
tool_creds.
Parameters
The function being decorated when used bare. Left as
None when the decorator is invoked with arguments.
Credential group(s) the tool needs access to. A single
group name, a list of group names, or None for no
credentials. Group names are configured in the active
Yera profile; the resolved keys appear as
<group>.<subkey> in
tool_creds.
Whether to append the tool's return value to the active
LLM context. When False, the Python caller still receives the
value, but it is not included in later model requests.
Returns
A DecoratedTool callable that resolves credentials and runs the function under the tool credential context when invoked. When called with arguments (fn is None), returns a decorator that produces the wrapper.
Raises
If creds is not None, a
str, or list[str].
Examples
Use the bare decorator for a tool with default behavior:
import yera as yr
@yr.tool
def celsius_to_fahrenheit(temperature: float) -> float:
return temperature * 9 / 5 + 32
Configure credentials and keep a sensitive result out of the model context:
import yera as yr
@yr.tool(creds="travel", insert_result=False)
def booking_headers() -> dict[str, str]:
token = yr.tool_creds().require("travel.token")
return {"Authorization": f"Bearer {token}"}