SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1610.05820 · 2016

Membership Inference Attacks Against Machine Learning Models

Vitaly Shmatikov, Reza Shokri, Marco Stronati

arXiv · PDF · Open in the Atlas

Code that ran

10 of 14 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 2 failures of our instrument; 4 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1610.05820.json · badge: /api/ran/1610.05820.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
Lasagne/Recipes canonical (classifier label)
classifier basis: census
1 ran · 2 no recorded run
PrivPkt/PrivPkt community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
yonsei-cysec/Membership_Inference_Attack community (classifier label)
classifier basis not recorded on the link
3 ran · 1 no recorded run
ml-research/do-not-trust-prediction-scores-for-membership-inference-attacks community (classifier label)
classifier basis not recorded on the link
1 ran · 0 no recorded run
spring-epfl/mia reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
accuracy Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yonsei-cysec/Membership_Inference_Attack/utils/metric.py
code served (permissive licence) · get_code("b0f936d4d6ae3b8c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned list, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create output tensor: batch_size=2, num_classes=4
# This represents logits from a model
output = torch.randn(2, 4)

# Create target tensor: batch_size=2
# These are class indices (0-3 for 4 classes)
target = torch.tensor([1, 3])

# topk parameter: tuple of k values to compute accuracy for
topk = (1, 2)

ARGS = {
    "output": output,
    "target": target,
    "topk": topk,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
force_length Ran
ran on a synthesized input
PrivPkt/PrivPkt/ISCX-preproc/kabab/utils/gen.py
code served (permissive licence) · get_code("5ec7cc132db252da")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Create a simple generator that yields some values
def sample_generator():
    for i in range(10):
        yield i

# Define ARGS dict with all parameters for force_length
ARGS = {
    "iterable": sample_generator(),
    "n": 5,
    "pad": 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_base_pkt Ran
ran on a synthesized input
PrivPkt/PrivPkt/ISCX-preproc/kabab/flow.py
code served (permissive licence) · get_code("9a16013f7255867e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# The function get_base_pkt indexes into cap with [0]
# So cap needs to be indexable (list, tuple, or tensor)
# We'll create a simple list/tuple of items

# Create a small capture object that can be indexed
cap = [
    torch.randn(2, 4),  # First element - this is what gets returned
    torch.randn(2, 4),  # Second element
    torch.randn(2, 4),  # Third element
]

ARGS = {
    "cap": cap
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_model_prediction_scores Ran
ran on a synthesized input
ml-research/do-not-trust-prediction-scores-for-membership-inference-attacks/utils/model_utils.py
code served (permissive licence) · get_code("af19edeb92a0ce40")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (4, 3) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn as nn

torch.manual_seed(0)

# Define a minimal model that satisfies the function's usage
class SimpleModel(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 3)
    
    def forward(self, x):
        return self.linear(x)

# Create model instance
model = SimpleModel()

# Create a minimal dataset (list of (x, y) tuples)
# x: batch of 4-dimensional inputs
# y: class labels (3 classes)
dataset = [
    (torch.randn(4), torch.tensor(0)),
    (torch.randn(4), torch.tensor(1)),
    (torch.randn(4), torch.tensor(2)),
    (torch.randn(4), torch.tensor(0)),
]

ARGS = {
    'model': model,
    'apply_softmax': True,
    'dataset': dataset,
    'batch_size': 2,
    'num_workers': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (4, 3) — shape (4, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_pcaps Ran
ran on a synthesized input
PrivPkt/PrivPkt/ISCX-preproc/kabab/utils/general.py
code served (permissive licence) · get_code("d16121d6ebb3bc1c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
from pathlib import Path


class MockPath:
    """Mock path object that mimics a pathlib-like interface with find_files()"""
    
    def __init__(self, base_dir):
        self.base_dir = base_dir
        self.files = []
    
    def add_file(self, filename):
        """Add a mock file to the directory"""
        self.files.append(MockFile(filename))
    
    def find_files(self):
        """Return list of mock files"""
        return self.files


class MockFile:
    """Mock file object with ext() method"""
    
    def __init__(self, filename):
        self.filename = filename
    
    def ext(self):
        """Return file extension without the dot"""
        if '.' in self.filename:
            return self.filename.split('.')[-1]
        return ''
    
    def __repr__(self):
        return f"MockFile({self.filename})"


# Create a mock data_dir with various file types
data_dir = MockPath("/mock/data")

# Add files with different extensions
data_dir.add_file("capture1.pcap")
data_dir.add_file("capture2.pcapng")
data_dir.add_file("config.json")
data_dir.add_file("results.csv")
data_dir.add_file("notes.txt")
data_dir.add_file("data.data")
data_dir.add_file("network_dump.pcap")
data_dir.add_file("stream.pcapng")

ARGS = {
    "data_dir": data_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
give_me_results Ran
ran on a synthesized input
PrivPkt/PrivPkt/mia/mia.py
code served (permissive licence) · get_code("d54e16609e4cd4a4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct arguments for give_me_results
# The function expects:
# - attack_guesses: binary predictions (0 or 1)
# - real_membership_labels: binary ground truth labels (0 or 1)

# Create small numpy arrays with binary values
attack_guesses = np.array([1, 0, 1, 1, 0, 0, 1, 0], dtype=np.int32)
real_membership_labels = np.array([1, 0, 1, 0, 0, 1, 1, 0], dtype=np.int32)

ARGS = {
    "attack_guesses": attack_guesses,
    "real_membership_labels": real_membership_labels,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
lira_metrics Ran
ran on a synthesized input
yonsei-cysec/Membership_Inference_Attack/utils/metric.py
code served (permissive licence) · get_code("7731cfa4bb088988")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# The function expects:
# - output: array-like of predicted labels (multi-label format)
# - target: array-like of true labels (multi-label format)
# - bool_average: bool
# - num_classes: int

# For multilabel_confusion_matrix, we need binary labels for each class
# Shape should be (n_samples, n_classes) with binary values

torch_seed = 0
np.random.seed(torch_seed)

# Create small multi-label data: 2 samples, 10 classes
# Each sample has binary labels for each class
n_samples = 2
n_classes = 10

# Generate random binary labels for target and output
target = np.random.randint(0, 2, size=(n_samples, n_classes))
output = np.random.randint(0, 2, size=(n_samples, n_classes))

ARGS = {
    "output": output,
    "target": target,
    "bool_average": True,
    "num_classes": n_classes,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_config Ran
ran on a synthesized input
yonsei-cysec/Membership_Inference_Attack/utils/load_config.py
code served (permissive licence) · get_code("ddfeb2a10eaefde7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned EasyDict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import yaml
from easydict import EasyDict
import os

# Create a temporary config file with YAML content
temp_dir = tempfile.mkdtemp()
config_path = os.path.join(temp_dir, 'config.yaml')

# Create a sample YAML config with multiple named configurations
config_data = {
    'model_config': {
        'learning_rate': 0.001,
        'batch_size': 32,
        'num_layers': 3,
        'hidden_dim': 128
    },
    'training_config': {
        'epochs': 100,
        'optimizer': 'adam',
        'loss': 'cross_entropy'
    },
    'default_config': {
        'seed': 42,
        'device': 'cpu'
    }
}

with open(config_path, 'w') as f:
    yaml.dump(config_data, f)

ARGS = {
    'config_name': 'model_config',
    'config_path': config_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
easydict, yaml — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EasyDict — a EasyDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a EasyDict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
prepare_attack_data Ran
ran; our contract draft was wrong, not the code
spring-epfl/mia/mia/estimators.py
code served (permissive licence) · get_code("f399cccee25a4712")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import sklearn.linear_model

# Create a simple classifier that has predict_proba method
class SimpleClassifier:
    def __init__(self):
        self.model = sklearn.linear_model.LogisticRegression(random_state=0, max_iter=1000)
        self.classes_ = np.array([0, 1])
        self.fitted = False
    
    def fit(self, X, y):
        self.model.fit(X, y)
        self.fitted = True
        return self
    
    def predict_proba(self, X):
        # Return probability estimates for 2 classes
        # Shape: (n_samples, 2)
        if not self.fitted:
            # Return dummy probabilities if not fitted
            n_samples = X.shape[0]
            probs = np.random.rand(n_samples, 2)
            probs = probs / probs.sum(axis=1, keepdims=True)
            return probs
        return self.model.predict_proba(X)

# Create the model
model = SimpleClassifier()

# Create training data (used in model)
X_in = np.random.rand(2, 4).astype(np.float32)  # 2 samples, 4 features
y_in = np.array([0, 1])  # 2 labels

# Create test data (not used in model training)
X_out = np.random.rand(3, 4).astype(np.float32)  # 3 samples, 4 features
y_out = np.array([1, 0, 1])  # 3 labels

# Prepare tuples
data_in = (X_in, y_in)
data_out = (X_out, y_out)

ARGS = {
    "model": model,
    "data_in": data_in,
    "data_out": data_out
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
preprocess Ran
ran on a synthesized input
Lasagne/Recipes/modelzoo/inception_v3.py
code served (permissive licence) · get_code("1dff06f483218f63")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 4, dtype family float, shape (1, 3, 299, 299) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
import torch

# Create a small 3-channel image with shape (height, width, channels)
# Using distinct dimensions: 5x7x3 so we can probe the shapes later
im = np.random.randint(0, 256, size=(5, 7, 3), dtype=np.uint8).astype(np.float32)

ARGS = {
    'im': im
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, skimage, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 4, dtype family float, shape (1, 3, 299, 299) — shape (1, 3, 299, 299) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
convert_back Not yet run
unverified (harvested, no recorded run)
Lasagne/Recipes/modelzoo/c3d.py
code served (permissive licence) · get_code("1ed624f64c635c68")
How it ran

no invocation recorded.

load_dataset Not yet run
unverified (harvested, no recorded run)
Lasagne/Recipes/papers/densenet/cifar10.py
code served (permissive licence) · get_code("5206fe60bb83055f")
How it ran

no invocation recorded.

load_pcap Not yet run
unverified (harvested, no recorded run)
PrivPkt/PrivPkt/ISCX-preproc/kabab/utils/general.py
code served (permissive licence) · get_code("21786ee5f33906cd")
How it ran

no invocation recorded.

make_member_nonmember Not yet run
unverified (harvested, no recorded run)
yonsei-cysec/Membership_Inference_Attack/shadow/make_data.py
code served (permissive licence) · get_code("d697fc8dd59df530")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 13 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
accuracy 116 other papers
1 dated earlier, 115 later · 8 only as an assembly brick
1502.03044 (2015-02)
Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
1905.03696 (2019-05)
HAWQ: Hessian AWare Quantization of Neural Networks with Mixed-Precision
1906.05849 (2019-06)
Contrastive Multiview Coding
and 113 more not named here

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

We quantitatively investigate how machine learning models leak information about the individual data records on which they were trained. We focus on the basic membership inference attack: given a data record and black-box access to a model, determine if the record was in the model's training dataset. To perform membership inference against a target model, we make adversarial use of machine learning and train our own inference model to recognize differences in the target model's predictions on the inputs that it trained on versus the inputs that it did not train on. We empirically evaluate our inference techniques on classification models trained by commercial "machine learning as a service" providers such as Google and Amazon. Using realistic datasets and classification tasks, including a hospital discharge dataset whose membership is sensitive from the privacy perspective, we show that these models can be vulnerable to membership inference attacks. We then investigate the factors that influence this leakage and evaluate mitigation strategies.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1610.05820")
get_code_for_paper("1610.05820")
have("1610.05820")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/1610.05820.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/1610.05820.svg)](https://syntology.ai/paper/1610.05820)

Connect an agent — have() is free.