SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.01489 · 2024

Agentless: Demystifying LLM-based Software Engineering Agents

arXiv · PDF · Open in the Atlas

Code that ran

15 of 24 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 11 have no contract verdict; 4 failures of our instrument; 9 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/2407.01489.json · badge: /api/ran/2407.01489.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
OpenAutoCoder/Agentless canonical (classifier label)
classifier basis: census
3 ran · 5 no recorded run
sorendunn/agentless-lite community (classifier label)
classifier basis not recorded on the link
8 ran · 3 no recorded run
experepair/experepair community (classifier label)
classifier basis not recorded on the link
4 ran · 1 no recorded run
FunctionStatusWhere it lives
check_python_syntax Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/syntax.py
code served (permissive licence) · get_code("3c6d16925eababcf")
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 subprocess
import tempfile
import os

# Create a temporary Python file with valid syntax
temp_file = tempfile.NamedTemporaryFile(mode='w', suffix='.py', delete=False)
temp_file.write("def hello():\n    print('Hello, World!')\n")
temp_file.close()

ARGS = {
    'filepath': temp_file.name,
    'timeout_seconds': 30
}
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 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
construct_file_meta_data Ran
ran; our contract draft was wrong, not the code
OpenAutoCoder/Agentless/agentless/fl/Index.py
code served (permissive licence) · get_code("4dc6308a692bee4f")
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 torch

ARGS = {
    'file_name': 'example_module.py',
    'clazzes': [
        {'name': 'MyClass'},
        {'name': 'AnotherClass'},
        {'name': 'ThirdClass'}
    ],
    'functions': [
        {'name': 'my_function'},
        {'name': 'another_function'},
        {'name': 'helper_func'}
    ]
}
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
count_and_organize_tasks Ran
ran on a synthesized input
experepair/experepair/ExpeRepair-v1.0/post_process.py
code served (permissive licence) · get_code("930c6e52cfa063f4")
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 str. 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 os
import shutil
import tempfile
from os.path import join as pjoin

# Create a temporary directory structure for testing
temp_dir = tempfile.mkdtemp()
expr_dir = temp_dir

# Create some dummy experiment result directories
task_exp_names = [
    "task_001_exp_a",
    "task_002_exp_b",
    "task_003_exp_c",
    "task_001_exp_d",
    "task_004_exp_e"
]

# Create these directories in the expr_dir
for exp_name in task_exp_names:
    os.makedirs(pjoin(expr_dir, exp_name), exist_ok=True)

# Define task lists
task_list = ["task_001", "task_002"]
task_list_name = "passed_tasks"

ARGS = {
    "task_list": task_list,
    "task_list_name": task_list_name,
    "task_exp_names": task_exp_names,
    "expr_dir": expr_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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
create_chatgpt_config Ran
ran; our contract draft was wrong, not the code
OpenAutoCoder/Agentless/agentless/util/api_requests.py
code served (permissive licence) · get_code("6aa11a092321f0eb")
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 torch
import numpy as np

# Construct valid arguments for create_chatgpt_config

# Test case 1: message as a string
message_str = "What is the capital of France?"
max_tokens_int = 100
temperature_float = 0.7
batch_size_int = 2
system_message_str = "You are a helpful assistant."
model_str = "gpt-3.5-turbo"

# Test case 2: message as a list of message dicts
message_list = [
    {"role": "user", "content": "Hello"},
    {"role": "assistant", "content": "Hi there!"},
    {"role": "user", "content": "How are you?"}
]

# Build ARGS with string message
ARGS = {
    "message": message_str,
    "max_tokens": max_tokens_int,
    "temperature": temperature_float,
    "batch_size": batch_size_int,
    "system_message": system_message_str,
    "model": model_str
}
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 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
extract_code_blocks Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/repair.py
code served (permissive licence) · get_code("70b519aa89fe249c")
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 re

def extract_code_blocks(text):
    pattern = '```(\\w+)\\n(.*?)\\n```'
    matches = re.findall(pattern, text, re.DOTALL)
    return [match[1] for match in matches]

# Setup code to construct valid arguments
text_with_code_blocks = """
Here is some documentation with code examples.


def hello_world():
    print("Hello, World!")


And here is another example:

javascript
console.log("Hello, World!");


Some more text here.

rust
fn main() {
    println!("Hello, World!");
}

"""

ARGS = {
    'text': text_with_code_blocks
}
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
find_consecutive_subset Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/utils.py
code served (permissive licence) · get_code("0f6b10ad5ce6a87b")
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)

# Analyze the function: find_consecutive_subset(numbers, x, no_start)
# - numbers: a list/iterable of numbers that gets sorted and deduplicated
# - x: an integer representing the length of consecutive subset to find
# - no_start: a collection (likely a set or list) of numbers to exclude from sequence start/end

# Create test data
numbers = [1, 5, 3, 2, 4, 6, 10, 9, 8]  # List with gaps and consecutive sequences
x = 3  # Looking for consecutive sequences of length 3
no_start = {10, 100}  # Numbers that shouldn't be at start or end of valid sequence

ARGS = {
    "numbers": numbers,
    "x": x,
    "no_start": no_start
}
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_processed_instances Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/utils.py
code served (permissive licence) · get_code("74352aad26882665")
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 set. 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 json
import tempfile
import os

# Create a temporary file with sample processed instances
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.jsonl', encoding='utf-8')
temp_file.write(json.dumps({'instance_id': 'id_1'}) + '\n')
temp_file.write(json.dumps({'instance_id': 'id_2'}) + '\n')
temp_file.write(json.dumps({'instance_id': 'id_3'}) + '\n')
temp_file.close()

ARGS = {
    'output_file': temp_file.name
}
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 set — a set output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a set 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
get_repo_folder_name Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/repo.py
code served (permissive licence) · get_code("5c56e3e454667f32")
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 str. 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.

# Setup code for get_repo_folder_name

ARGS = {
    "repo_url": "https://github.com/user/my-repo-name"
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
load_json Ran
ran on a synthesized input
experepair/experepair/ExpeRepair-v1.0/agentless_utils.py
code served (permissive licence) · get_code("227736b3e5c52f62")
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 json
import tempfile
import os

# Create a temporary JSON file with valid content
temp_dir = tempfile.mkdtemp()
temp_json_file = os.path.join(temp_dir, "test.json")

test_data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": {"nested": "object"}
}

with open(temp_json_file, 'w') as f:
    json.dump(test_data, f)

ARGS = {
    "filepath": temp_json_file
}
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 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
load_jsonl Ran
ran on a synthesized input
experepair/experepair/ExpeRepair-v1.0/agentless_utils.py
code served (permissive licence) · get_code("62a5fcf064a81c58")
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 json
import tempfile
import os

# Create a temporary JSONL file with test data
temp_dir = tempfile.mkdtemp()
temp_filepath = os.path.join(temp_dir, "test.jsonl")

# Write test JSONL data
test_data = [
    {"id": 1, "text": "hello world"},
    {"id": 2, "text": "foo bar"},
    {"id": 3, "name": "test", "value": 42}
]

with open(temp_filepath, 'w') as f:
    for item in test_data:
        f.write(json.dumps(item) + '\n')

ARGS = {
    "filepath": temp_filepath
}
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
parse_edit_command Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/repair.py
code served (permissive licence) · get_code("68f6c6922f31e7e1")
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.

# Setup code for parse_edit_command

ARGS = {
    'edit_string': """# myfile.py
< SEARCH
def hello():
    print("old")
==================================================
> REPLACE
def hello():
    print("new")
"""
}
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 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
remove_lines Ran
ran; our contract draft was wrong, not the code
OpenAutoCoder/Agentless/agentless/util/compress_file.py
code served (permissive licence) · get_code("a139d312ce4671dd")
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 str. 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.

raw_code = """def hello():
    print("Hello, World!")
    x = 1 + 2
    result = x * 5
    return result

def goodbye():
    print("Goodbye!")
    return None"""

remove_line_intervals = [(2, 3), (6, 7)]

ARGS = {
    "raw_code": raw_code,
    "remove_line_intervals": remove_line_intervals,
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
remove_outer_folder Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/repo.py
code served (permissive licence) · get_code("fd177a088c798c08")
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 str. 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 os

def remove_outer_folder(path):
    _, remaining_path = os.path.split(os.path.dirname(path))
    result = os.path.join(remaining_path, os.path.basename(path))
    return result

# The function takes a file path and removes the outermost folder from it
# For example: /a/b/c/file.txt -> b/c/file.txt

ARGS = {
    "path": "/home/user/documents/file.txt"
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
replace_html_tags Ran
ran; our contract draft was wrong, not the code
experepair/experepair/ExpeRepair-v1.0/log.py
pointer only (licence: MIT); the stored copy get_code serves, from GAIR-NLP/OpenSWE, is pointer only (licence: NOASSERTION) · get_code("07299a3c29c33ec8")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from GAIR-NLP/OpenSWE, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

setup_logging Ran
ran on a synthesized input
sorendunn/agentless-lite/agentless_lite/util/logging.py
code served (permissive licence) · get_code("c9355b8fec2e5c50")
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 Logger. 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 logging
import os
import tempfile

# Create a temporary directory for logs
log_dir = tempfile.mkdtemp()

ARGS = {
    'instance_id': 'test_instance_001',
    'log_dir': log_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 Logger — a Logger output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Logger 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
combine_by_instance_id Not yet run
unverified (harvested, no recorded run)
experepair/experepair/ExpeRepair-v1.0/agentless_utils.py
code served (permissive licence) · get_code("f8732f0644352106")
How it ran

no invocation recorded.

get_all_patches Not yet run
unverified (harvested, no recorded run)
OpenAutoCoder/Agentless/agentless/repair/rerank.py
code served (permissive licence) · get_code("502ce1b1b7bb81e3")
How it ran

no invocation recorded.

get_all_patches_num Not yet run
unverified (harvested, no recorded run)
OpenAutoCoder/Agentless/agentless/repair/rerank.py
code served (permissive licence) · get_code("2fb4111f1cf4cd40")
How it ran

no invocation recorded.

get_base64_from_url Not yet run
unverified (harvested, no recorded run)
sorendunn/agentless-lite/agentless_lite/util/backends.py
code served (permissive licence) · get_code("c8bcfbded40b94b2")
How it ran

no invocation recorded.

get_sample Not yet run
unverified (harvested, no recorded run)
OpenAutoCoder/Agentless/agentless/repair/rerank.py
code served (permissive licence) · get_code("6f12cc79072e9048")
How it ran

no invocation recorded.

num_tokens_from_messages Not yet run
unverified (harvested, no recorded run)
OpenAutoCoder/Agentless/agentless/util/api_requests.py
code served (permissive licence) · get_code("2df3e1e980b47ac5")
How it ran

no invocation recorded.

num_tokens_from_messages Not yet run
unverified (harvested, no recorded run)
sorendunn/agentless-lite/agentless_lite/util/repair.py
code served (permissive licence) · get_code("6475d4ef21dc4237")
How it ran

no invocation recorded.

request_chatgpt_engine Not yet run
unverified (harvested, no recorded run)
OpenAutoCoder/Agentless/agentless/util/api_requests.py
code served (permissive licence) · get_code("aabc18aa41c3a42b")
How it ran

no invocation recorded.

request_chatgpt_engine Not yet run
unverified (harvested, no recorded run)
sorendunn/agentless-lite/agentless_lite/util/backends.py
code served (permissive licence) · get_code("71476abc302492f9")
How it ran

no invocation recorded.

Also present in other papers' code

Of 24 samples lifted from this paper's linked repositories, 24 carry a function identity and 0 do not. Of the 24 distinct functions with an identity, 4 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 2 later) and 20 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
replace_html_tags 2 other papers
1 dated earlier, 1 later
2404.05427 (2024-04)
AutoCodeRover: Autonomous Program Improvement
2603.13023 (2026-03)
daVinci-Env: Open SWE Environment Synthesis at Scale
construct_file_meta_data 1 other paper
1 later
2502.19811 (2025-02)
Comet: Fine-grained Computation-communication Overlapping for Mixture-of-Experts
create_chatgpt_config 1 other paper
1 dated earlier
2403.19114 (2024-03)
Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM
num_tokens_from_messages 1 other paper
1 dated earlier
2403.19114 (2024-03)
Top Leaderboard Ranking = Top Coding Proficiency, Always? EvoEval: Evolving Coding Benchmarks via LLM

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

Recent advancements in large language models (LLMs) have significantly advanced the automation of software development tasks, including code synthesis, program repair, and test generation. More recently, researchers and industry practitioners have developed various autonomous LLM agents to perform end-to-end software development tasks. These agents are equipped with the ability to use tools, run commands, observe feedback from the environment, and plan for future actions. However, the complexity of these agent-based approaches, together with the limited abilities of current LLMs, raises the following question: Do we really have to employ complex autonomous software agents? To attempt to answer this question, we build Agentless -- an agentless approach to automatically solve software development problems. Compared to the verbose and complex setup of agent-based approaches, Agentless employs a simplistic three-phase process of localization, repair, and patch validation, without letting the LLM decide future actions or operate with complex tools. Our results on the popular SWE-bench Lite benchmark show that surprisingly the simplistic Agentless is able to achieve both the highest performance (32.00%, 96 correct fixes) and low cost ($0.70) compared with all existing open-source software agents! Furthermore, we manually classified the problems in SWE-bench Lite and found problems with exact ground truth patch or insufficient/misleading issue descriptions. As such, we construct SWE-bench Lite-S by excluding such problematic issues to perform more rigorous evaluation and comparison. Our work highlights the current overlooked potential of a simple, interpretable technique in autonomous software development. We hope Agentless will help reset the baseline, starting point, and horizon for autonomous software agents, and inspire future work along this crucial direction.

For agents

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

get_harvested_code_for_paper("2407.01489")
get_code_for_paper("2407.01489")
have("2407.01489")

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

curl https://syntology.ai/api/ran/2407.01489.json

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

[![Syntology run record](https://syntology.ai/api/ran/2407.01489.svg)](https://syntology.ai/paper/2407.01489)

Connect an agent — have() is free.