SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2603.15925 · ICML · 2026

Generative Inverse Design with Abstention via Diagonal Flow Matching

Hanno Gottschalk, Miguel De Campos, Werner Krebs

arXiv · PDF · Open in the Atlas

Code that ran

12 of 15 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; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 11 failures of our instrument; 3 no recorded run), measured 2026-11-22; 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/2603.15925.json · badge: /api/ran/2603.15925.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
migueldecampos/diagcfm canonical (classifier label)
classifier basis: census
12 ran · 3 no recorded run
FunctionStatusWhere it lives
augment_design Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/diagcfm_core.py
code served (permissive licence) · get_code("14579f6fd357c576")
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, 7) 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

torch.manual_seed(0)

# Create a small input tensor: batch_size=2, feature_dim=4
x = torch.randn(2, 4)

# Set number of labels
num_labels = 3

ARGS = {
    "x": x,
    "num_labels": num_labels,
}
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 (2, 7) — shape (2, 7) 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
augment_labels Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/diagcfm_core.py
code served (permissive licence) · get_code("a950a96fae09d6ba")
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, 7) 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

torch.manual_seed(0)

# y is a tensor with shape (batch_size, num_labels)
# num_design_params is an int specifying the number of design parameters
y = torch.randn(2, 4)  # batch_size=2, num_labels=4
num_design_params = 3

ARGS = {
    "y": y,
    "num_design_params": num_design_params,
}
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 (2, 7) — shape (2, 7) 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
binary_auroc Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
migueldecampos/diagcfm/benchmarking.py
code served (permissive licence) · get_code("291cbf5097332827")
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 float, ndim 0, dtype family float, shape () 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

# Create simple test data for binary AUROC
# scores: predicted scores/probabilities
# labels: binary labels (0 or 1)

# Create a small example with some clear separation
scores = np.array([0.1, 0.2, 0.7, 0.8, 0.9, 0.3], dtype=np.float64)
labels = np.array([0, 0, 1, 1, 1, 0], dtype=bool)

ARGS = {
    'scores': scores,
    'labels': 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 float, ndim 0, dtype family float, shape () — shape () 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
checkpoint_args Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/training.py
code served (permissive licence) · get_code("9256d0d79fd9454f")
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 dict. 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 argparse
from pathlib import Path
import torch
import torch.nn as nn

torch.manual_seed(0)

# Create a minimal argparse.Namespace with various types of values
# that would typically be passed to checkpoint_args
args = argparse.Namespace(
    model_name="bert-base",
    learning_rate=0.001,
    num_epochs=10,
    batch_size=32,
    checkpoint_path=Path("/tmp/model.pt"),
    output_dir=Path("/home/user/results"),
    seed=42,
    use_cuda=False,
    config_dict={"hidden_size": 768, "num_layers": 12},
    tensor_value=torch.randn(2, 4),
    list_value=[1, 2, 3],
    float_value=3.14159,
    int_value=100,
    bool_value=True,
    string_value="test_run"
)

ARGS = {"args": args}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
compute_image_stats Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/fashionmnist_stats.py
code served (permissive licence) · get_code("f2e2a80f71a66a5a")
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, 5) 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 torch.nn.functional as F

torch.manual_seed(0)

# Create a small batch of images
# Shape: (batch_size, channels, height, width)
# Using distinct sizes: batch=2, channels=3, height=4, width=5
images = torch.randn(2, 3, 4, 5)

ARGS = {
    "images": images,
}
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 (2, 5) — shape (2, 5) 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
dtlz2_forward Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
migueldecampos/diagcfm/dtlz2.py
code served (permissive licence) · get_code("225843d203c0fe70")
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, 3) 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 math

torch.manual_seed(0)

# Create a valid input tensor with shape (batch_size, num_design_params)
# batch_size = 2, num_design_params = 4
x = torch.rand(2, 4, dtype=torch.float32)

# num_objectives must be <= num_design_params
# Setting num_objectives = 3 (which is valid since 3 <= 4)
num_objectives = 3

ARGS = {
    "x": x,
    "num_objectives": num_objectives,
}
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 (2, 3) — shape (2, 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
fashionmnist_forward Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/fashionmnist_stats.py
code served (permissive licence) · get_code("b72d6e7bd83989ed")
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, 5) 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

torch.manual_seed(0)

# The function fashionmnist_forward expects:
# - x: torch.Tensor that will be reshaped to (-1, 1, 28, 28)
# 
# Since IMAGE_SHAPE = (1, 28, 28), the input x should be a flattened version
# of images with shape (batch_size, 1*28*28) = (batch_size, 784)
#
# Let's create a small batch of 2 flattened images

batch_size = 2
flattened_dim = 1 * 28 * 28  # 784

# Create a tensor with shape (2, 784) representing 2 flattened images
x = torch.randn(batch_size, flattened_dim)

ARGS = {
    "x": x
}
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 (2, 5) — shape (2, 5) 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_device Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/diagcfm_core.py
code served (permissive licence) · get_code("900aec308656ff63")
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 device. 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 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)

ARGS = {
    'name': 'cpu'
}
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 device — a device output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a device 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
label_mse Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/benchmarking.py
code served (permissive licence) · get_code("9fb31a9c2bbcfdeb")
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 1, dtype family float, shape (2,) 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 torch.nn as nn

torch.manual_seed(0)

# Build a simple forward function that takes a tensor and returns predictions
class SimpleModel(nn.Module):
    def __init__(self, input_dim, output_dim):
        super().__init__()
        self.linear = nn.Linear(input_dim, output_dim)
    
    def forward(self, x):
        return self.linear(x)

# Create model instance
model = SimpleModel(input_dim=4, output_dim=3)

# Create forward_fn that wraps the model
def forward_fn(designs):
    return model(designs)

# Test case 1: 2D designs (B, P)
designs_2d = torch.randn(2, 4)
target_labels_2d = torch.randn(2, 3)

# Test case 2: 3D designs (K, B, P)
designs_3d = torch.randn(5, 2, 4)
target_labels_3d = torch.randn(2, 3)

# Use 2D case for the arguments
ARGS = {
    "designs": designs_2d,
    "target_labels": target_labels_2d,
    "forward_fn": forward_fn
}
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 1, dtype family float, shape (2,) — shape (2,) 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
make_dtlz2_loaders Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/dtlz2.py
code served (permissive licence) · get_code("90269d97711437e6")
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 torch
import numpy as np

torch.manual_seed(0)
np.random.seed(0)

ARGS = {
    "num_design_params": 5,
    "num_objectives": 3,
    "train_samples": 2,
    "val_samples": 2,
    "g_max": 0.5,
    "batch_size": 2,
    "seed": 42,
}
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 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
sample_dtlz2_stratified Ran
ran, honoured the contract we drafted
migueldecampos/diagcfm/dtlz2.py
code served (permissive licence) · get_code("0af183f4648157dc")
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, 5) 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

torch.manual_seed(0)

ARGS = {
    "num_samples": 2,
    "num_design_params": 5,
    "num_objectives": 3,
    "g_max": 1.5,
    "seed": 42,
}
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, 5) — shape (2, 5) 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
tmm_reflectance Ran
ran; our contract draft was wrong, not the code
migueldecampos/diagcfm/thin_film.py
code served (permissive licence) · get_code("dec14035f4e6da15")
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, 3) 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

torch.manual_seed(0)

# Parameters for tmm_reflectance
batch_size = 2
num_layers = 4
num_wavelengths = 3

# thickness_nm: shape (batch_size, num_layers)
thickness_nm = torch.rand(batch_size, num_layers) * 100 + 50

# refractive_indices: shape (num_layers,)
refractive_indices = torch.tensor([1.5, 2.0, 1.8, 1.6], dtype=torch.float32)

# wavelength_nm: shape (num_wavelengths,)
wavelength_nm = torch.tensor([400.0, 550.0, 700.0], dtype=torch.float32)

ARGS = {
    "thickness_nm": thickness_nm,
    "refractive_indices": refractive_indices,
    "wavelength_nm": wavelength_nm,
}
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 (2, 3) — shape (2, 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
layer_indices Not yet run
unverified (harvested, no recorded run)
migueldecampos/diagcfm/thin_film.py
code served (permissive licence) · get_code("2f0487e7da16f7f3")
How it ran

no invocation recorded.

make_fashionmnist_loaders Not yet run
unverified (harvested, no recorded run)
migueldecampos/diagcfm/fashionmnist_stats.py
code served (permissive licence) · get_code("3e51f30ab97f29eb")
How it ran

no invocation recorded.

wavelengths Not yet run
unverified (harvested, no recorded run)
migueldecampos/diagcfm/thin_film.py
code served (permissive licence) · get_code("23bf56b8e6a93b53")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 15 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.

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

Inverse design aims to find design parameters x achieving target performance y * . Generative approaches learn bidirectional mappings between designs and labels, enabling diverse solution sampling. However, standard conditional flow matching (CFM), when adapted to inverse problems by pairing labels with design parameters, exhibits strong sensitivity to their arbitrary ordering and scaling, leading to unstable training. We introduce Diagonal Flow Matching (Diag-CFM), which resolves this through a zero-anchoring strategy that pairs design coordinates with noise and labels with zero, making the learning problem provably invariant to coordinate permutations. This yields substantially lower round-trip error than CFM and invertible neural network baselines across design dimensions up to P =784, including orderof-magnitude gains on several benchmarks. We develop two architecture-intrinsic uncertainty metrics, Zero-Deviation and Self-Consistency, that enable three practical capabilities: selecting the best candidate among multiple generations, abstaining from unreliable predictions, and detecting out-ofdistribution targets; consistently outperforming ensemble and general-purpose alternatives across all tasks. We validate on airfoil, gas turbine combustor, scalable analytical benchmarks, a photonics inverse-design task, and an image-statistics benchmark.

For agents

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

get_harvested_code_for_paper("2603.15925")
get_code_for_paper("2603.15925")
have("2603.15925")

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

curl https://syntology.ai/api/ran/2603.15925.json

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

[![Syntology run record](https://syntology.ai/api/ran/2603.15925.svg)](https://syntology.ai/paper/2603.15925)

Connect an agent — have() is free.