Core

class aisploit.core.BaseCallbackHandler

Bases: object

Base class for callback handlers.

on_redteam_attempt_end(attempt: int, response: Response, score: Score, *, run_id: str)

Called when a red team attempt ends.

Args:

attempt (int): The attempt number. response (Response): The response from the attempt. score (Score): The score of the attempt. run_id (str): The ID of the current run.

on_redteam_attempt_start(attempt: int, prompt: PromptValue, *, run_id: str)

Called when a red team attempt starts.

Args:

attempt (int): The attempt number. prompt (BasePromptValue): The prompt value. run_id (str): The ID of the current run.

on_scanner_plugin_end(name: str, *, run_id: str)

Called when a scanner plugin ends.

Args:

name (str): The name of the scanner plugin. run_id (str): The ID of the current run.

on_scanner_plugin_start(name: str, *, run_id: str)

Called when a scanner plugin starts.

Args:

name (str): The name of the scanner plugin. run_id (str): The ID of the current run.

on_sender_send_prompt_end()

on_sender_send_prompt_start()

class aisploit.core.BaseChatModel

Bases: Runnable[PromptValue | str | Sequence[BaseMessage | List[str] | Tuple[str, str] | str | Dict[str, Any]], BaseMessage]

abstract supports_functions() → bool

Check if the model supports additional functions beyond basic chat.

Returns:

bool: True if the model supports additional functions, False otherwise.

class aisploit.core.BaseChatModelConverter(chat_model: aisploit.core.model.BaseChatModel)

Bases: BaseConverter, ABC

chat_model: BaseChatModel

class aisploit.core.BaseClassifier

Bases: ABC, Generic[T, Input]

An abstract base class for classifiers.

abstract score(input: Input, references: List[Input] | None = None, metadata: Dict[str, Any] | None = None) → Score[T]

Score the input and return a Score object.

Args:

input (Input): The input to be scored. references (List[Input], optional): List of reference inputs. Defaults to None. metadata (Dict[str, Any], optional): Additional metadata for scoring. Defaults to {}.

Returns:

Score[T]: A Score object representing the score of the input.

tags: List[str]

class aisploit.core.BaseConverter

Bases: ABC

Abstract base class for converters.

convert(prompt: str | PromptValue) → PromptValue

Converts the prompt value.

Args:

prompt (BasePromptValue): The prompt value to be converted.

Returns:

BasePromptValue: The converted prompt value.

class aisploit.core.BaseDataset

Bases: ABC

Generic dataset class.

aisploit.core.BaseEmbeddings

alias of Embeddings

class aisploit.core.BaseGenerator

Bases: Generic[T], ABC

abstract generate() → Generator[T, Any, None]

class aisploit.core.BaseImageTarget(size: Literal['512x512', '1024x1024'] = '512x512', show_image: bool = False)

Bases: ABC

show_image: bool

size: Literal['512x512', '1024x1024']

class aisploit.core.BaseJob(*, disable_progressbar: bool = <factory>, verbose: bool = False)

Bases: object

disable_progressbar: bool

verbose: bool

class aisploit.core.BaseLLM

Bases: Runnable[PromptValue | str | Sequence[BaseMessage | List[str] | Tuple[str, str] | str | Dict[str, Any]], str]

aisploit.core.BasePromptValue

alias of PromptValue

class aisploit.core.BaseReport(*, run_id: str)

Bases: Generic[T], ABC

A generic base class for reports.

property created_at: datetime

class aisploit.core.BaseTarget

Bases: ABC

An abstract base class for targets.

abstract send_prompt(prompt: PromptValue) → Response

Send a prompt to the target and return the response.

Args:

prompt (BasePromptValue): The prompt to send.

Returns:

Response: The response from the target.

class aisploit.core.BaseTextClassifier

Bases: BaseClassifier[T, str], Generic[T]

An abstract base class for text classifiers.

aisploit.core.BaseVectorStore

alias of VectorStore

class aisploit.core.CallbackManager(*, run_id: str, callbacks: Sequence[BaseCallbackHandler] | None)

Bases: object

Manages callback handlers.

on_redteam_attempt_end(attempt: int, response: Response, score: Score)

Notify callback handlers when a red team attempt ends.

Args:

attempt (int): The attempt number. response (Response): The response from the attempt. score (Score): The score of the attempt.

on_redteam_attempt_start(attempt: int, prompt: PromptValue)

Notify callback handlers when a red team attempt starts.

Args:

attempt (int): The attempt number. prompt (BasePromptValue): The prompt value.

on_scanner_plugin_end(name: str)

Notify callback handlers when a scanner plugin ends.

Args:

name (str): The name of the scanner plugin.

on_scanner_plugin_start(name: str)

Notify callback handlers when a scanner plugin starts.

Args:

name (str): The name of the scanner plugin.

on_sender_send_prompt_end()

on_sender_send_prompt_start()

exception aisploit.core.ContentFilteredException

Bases: Exception

class aisploit.core.DataclassDataset

Bases: BaseDataset, Generic[T]

Dataset class based on dataclasses.

class aisploit.core.Prompt(value: str | langchain_core.prompt_values.PromptValue, metadata: Dict[str, Any] = <factory>)

Bases: object

metadata: Dict[str, Any]

value: str | PromptValue

class aisploit.core.Response(content: str, metadata: ~typing.Dict[str, ~typing.Any] = <factory>)

Bases: object

A class representing a response from the target.

content: str

metadata: Dict[str, Any]

class aisploit.core.Score(flagged: bool, value: T, description: str = '', explanation: str = '')

Bases: Generic[T]

A class representing a score.

Attributes:

flagged (bool): Whether the score is flagged. value (T): The value of the score. description (str): Optional description of the score. explanation (str): Optional explanation of the score.

description: str

explanation: str

flagged: bool

value: T

class aisploit.core.StringPromptValue(*, text: str, type: Literal['StringPromptValue'] = 'StringPromptValue')

Bases: PromptValue

String prompt value.

classmethod get_lc_namespace() → List[str]

Get the namespace of the langchain object.

text: str

Prompt text.

to_messages() → List[BaseMessage]

Return prompt as messages.

to_string() → str

Return prompt as string.

type: Literal['StringPromptValue']

class aisploit.core.TabularDataset

Bases: BaseDataset

Dataset class for tabular data.

class aisploit.core.YamlDeserializable

Bases: object

Base class for objects that can be deserialized from YAML.

classmethod from_yaml_file(file: Path) → T

Load an object from a YAML file.

Args:

cls (Type[T]): The class to instantiate. file (Path): The path to the YAML file.

Returns:

T: An instance of the class deserialized from the YAML file.

Raises:

FileNotFoundError: If the specified file does not exist. ValueError: If there’s an error in parsing the YAML data.