SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.14849 · 2026

Atomix: Timely, Transactional Tool Use for Reliable Agentic Workflows

Lars Klein, Nearchos Potamitis, Akhil Arora, Laurent Bindschaedler, Bardia Mohammadi

arXiv · PDF · Open in the Atlas

Code that ran

3 of 6 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 0 have no contract verdict; 3 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/2602.14849.json · badge: /api/ran/2602.14849.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
mpi-dsg/atomix canonical (classifier label)
classifier basis: census
3 ran · 3 no recorded run
FunctionStatusWhere it lives
get_logger Ran
ran; our contract draft was wrong, not the code
mpi-dsg/atomix/src/atomix/logging.py
code served (permissive licence) · get_code("5c22288de7d5c59c")
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

# The function get_logger takes a single parameter: name (str)
# The function checks if the name starts with 'atomix.' and prepends it if not
# Then returns a logger instance using logging.getLogger()

# Create a simple test case
ARGS = {
    'name': 'my_module'
}
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
parse_flags Ran
ran; our contract draft was wrong, not the code
mpi-dsg/atomix/src/atomix/config.py
code served (permissive licence) · get_code("b53d4f5e18d4ad4a")
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 AblationFlags. 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 parse_flags function

ARGS = {
    "spec": "Tx-NoScopeOnRead+Tx-GlobalFrontier"
}
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 AblationFlags — a AblationFlags output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a AblationFlags 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
setup_logging Ran
ran; our contract draft was wrong, not the code
mpi-dsg/atomix/src/atomix/logging.py
code served (permissive licence) · get_code("2da8052ea7b493ae")
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 sys
from io import StringIO

# Set up basic logging to avoid conflicts
logging.basicConfig(level=logging.DEBUG, force=True)

# Create a StringIO stream to capture output
test_stream = StringIO()

ARGS = {
    'level': logging.DEBUG,
    'format_string': '%(asctime)s [%(levelname)s] %(name)s: %(message)s',
    'stream': test_stream,
}
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
cmd_demo Not yet run
unverified (harvested, no recorded run)
mpi-dsg/atomix/src/atomix/__main__.py
code served (permissive licence) · get_code("ca6a2fb19281dd0c")
How it ran

no invocation recorded.

cmd_osworld Not yet run
unverified (harvested, no recorded run)
mpi-dsg/atomix/src/atomix/__main__.py
code served (permissive licence) · get_code("a78220dd7371f58b")
How it ran

no invocation recorded.

cmd_version Not yet run
unverified (harvested, no recorded run)
mpi-dsg/atomix/src/atomix/__main__.py
code served (permissive licence) · get_code("88752738b11473da")
How it ran

no invocation recorded.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 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 6 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

LLM agents execute multi-step workflows that mutate external state through tools. Common orchestrators treat tool return as the settlement trigger, so faults, speculation, and concurrent agents can leave partial effects, losing-branch residue, stale writes, or irreversible sends. Correct settlement needs two facts that retries, checkpoint replay, locks, and compensation each conflate: which effects must settle together, and when earlier conflicting work is exhausted. Atomix makes this split explicit with progress-aware transactions. The runtime records reads and effects during execution, seals a transaction when its footprint is complete, and commits only after per-resource frontiers show that no earlier conflicting work can still arrive. Commit is final settlement: Atomix releases bufferable effects, accepts reversible external effects as final, and lets irreversible effects leave the gate. Abort suppresses unreleased effects and compensates externalized reversible effects where possible. On representative agent workloads, this composition improves clean recovery under injected faults, isolates contending and speculative work, and prevents correctly classified irreversible actions from leaking; microbenchmarks show microsecond-scale wrapper overhead relative to tool latency. The code and artifacts of Atomix are available at https://github.com/mpi-dsg/atomix. 1 Introduction LLM agents execute multi-step tasks that mutate external state: editing code, updating databases, filling web forms, sending emails, and calling third-party APIs LangChain Inc. [2026], CrewAI, Inc. [2026], Wu et al. [2024], Anthropic [2026], OpenAI [2026], Model Context Protocol Project [2025]. These workloads increasingly use concurrent execution: multiple agents share resources, parallel plans explore alternatives, and speculative execution starts downstream work before upstream confirmation. Each pattern raises the same question: when may a tool effect become permanent? Current agent frameworks expose this decision indirectly. Orchestrators provide checkpointing, retries, and timeouts, and rely on tool-level idempotency keys plus Saga-style compensations LangChain Inc. [2026], Temporal Technologies Inc. [2026], Amazon Web Services [2026], Garcia-Molina and Salem [1987]. These mechanisms recover after a tool effect has externalized or after the workflow restores its own state. That execution model leaves three gaps once workflows run in parallel: 1. Speculation: parallel plans execute real tool calls; losing branches leave residual side effects. 2. Contention: concurrent agents touch the same resource and interleave writes. 3. Irreversibility: some effects cannot be undone once externalized. Atomix adds transactional settlement at the tool interface. Atomix separates execution, sealing, frontier checks, and settlement. During execution, adapters record the resources a transaction reads and the effects it may produce. The orchestrator then seals the transaction, declaring that no more reads, scopes, or effects will be added. Atomix commits a sealed transaction only when per-resource frontiers certify, under the orchestrator's advancement rule, that no earlier work on any Preprint.

For agents

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

get_harvested_code_for_paper("2602.14849")
get_code_for_paper("2602.14849")
have("2602.14849")

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

curl https://syntology.ai/api/ran/2602.14849.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.14849.svg)](https://syntology.ai/paper/2602.14849)

Connect an agent — have() is free.