system stringclasses 211
values | prompt stringlengths 245 106k | ground_truth stringlengths 34 30.3k | domain stringclasses 7
values |
|---|---|---|---|
Role: Operate as a language world model for a software-engineering shell testbed. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. The curr... | **Turn 1 User**
###[Task Description]
Crash when unpacking non TypeVarTuple in args list to `Callable` alias
<!--
Use this form only if mypy reports an "INTERNAL ERROR" and/or gives a traceback.
Please include the traceback and all other messages below (use `mypy --show-traceback`).
-->
**Crash Report**
... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": " if isinstance(repl, TypeVarTupleType):\n return repl\n raise NotImplementedError\n\n def visit_unpack_type(self, t: UnpackType) -> Type:\n # It is impossible to reasonably implement visit_unpack_type, because\n ... | swe |
Role: You are a language world model for a web search and page-browsing environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and an... | **Turn 1 User**
###[Goal]
Find information about what WNZK is and the city where it is located. WNZK is likely a radio station or other type of location identifier.
###[Action]
Use `browser.search` to search the web or retrieval index for relevant documents.
- This advances the task or changes state.
- Args: `query`="W... | ###[Observation]
Doc 220
URL: https://radio-locator.com/info/WNZK-AM
L1: WNZK-AM 680 kHz
L2:
L3: Dearborn Heights, Michigan
L4: "Station of the Nations"
L5:
L6:
L7: Website:
L8: https://www.birach.com/wnzk.html
L9:
L10: Audio Feed:
L11: http://wnzk.birach.com:9000/stream
L12:
L13: Station Owner:
L14: Birach Broadc... | search |
Role: Act as a language world model for a web search and page-browsing environment. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, and prior obse... | **Turn 1 User**
###[Goal]
How many counties are there in the largest state in the region where the film The Family Stone was set? You should follow the format instruction in the requestion strictly and wrap the final answer in \boxed{}.
###[Action]
Use `browser.search` to search the web or retrieval index for relevant ... | ###[Observation]
https://en.wikipedia.org/wiki/The_Family_Stone
URL: https://en.wikipedia.org/wiki/The_Family_Stone
L1: Document not found | search |
Role: Act as a language world model for a terminal command-execution environment. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, and prior observ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a platform engineer responsible for maintaining our CI/CD pipelines. We are building a new artifact validation system.
There are two components to this system:
1. A small Rust CLI tool used by build agents to compute Adler-32 checksums of artifacts.
2. A Pyth... | ###[Observation]
[package]
name = "rust_helper"
version = "0.1.0"
edition = "2021"
[dependencies]
adler = "1.0.2"
(exit_code=0) | terminal |
Role: You are a language world model for a desktop accessibility-control environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and a... | **Turn 1 User**
###[Observation]
Computer 'Program Manager', b=0,0,1920,1080, raw_control='PaneControl'
node=1 generic '', raw_control='PaneControl', b=0,0,1920,1080
node=2 list 'Desktop', raw_control='ListControl', automation_id='1', b=0,0,1920,1080
node=3 listitem 'Recycle Bin', raw_control='ListItemControl', b... | ###[Observation]
Computer 'New Tab - Google Chrome', b=15,15,1590,993, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=26,15,1579,982
node=2 generic 'New Tab - Google Chrome', raw_control='PaneControl', b=25,15,1580,983
node=3 generic '', raw_control='PaneControl', b=25,15,1580,983
no... | os |
Role: Act as a language world model for a browser accessibility environment with reference IDs. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, an... | **Turn 1 User**
###[Observation]
- RootWebArea "(2 unread) - webtasks@yahoo.com - Yahoo Mail" [ref=e1]
- banner [ref=e2]
- banner [ref=e3]
- navigation [ref=e4]
- list [ref=e5]
- listitem [ref=e6]
- link " HOME" [ref=e7]
- graphics-symbol [ref=e8]
- list... | ###[Observation]
- RootWebArea "(2 unread) - webtasks@yahoo.com - Yahoo Mail" [ref=e1]
- banner [ref=e2]
- banner [ref=e3]
- navigation [ref=e4]
- list [ref=e5]
- listitem [ref=e6]
- link " HOME" [ref=e7]
- graphics-symbol [ref=e8]
- listitem [ref=e9]
... | web |
Role: Act as a language world model for a browser accessibility environment with reference IDs. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, an... | **Turn 1 User**
###[Observation]
- RootWebArea "(2 unread) - webtasks@yahoo.com - Yahoo Mail" [ref=e1]
- banner [ref=e2]
- banner [ref=e3]
- navigation [ref=e4]
- list [ref=e5]
- listitem [ref=e6]
- link " HOME" [ref=e7]
- graphics-symbol [ref=e8]
- list... | ###[Observation]
- RootWebArea "(2 unread) - webtasks@yahoo.com - Yahoo Mail" [ref=e1]
- banner [ref=e2]
- banner [ref=e3]
- navigation [ref=e4]
- list [ref=e5]
- listitem [ref=e6]
- link " HOME" [ref=e7]
- graphics-symbol [ref=e8]
- listitem [ref=e9]
... | web |
Role: You are a state-transition world model for a web search and page-browsing environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. Apply the ... | **Turn 1 User**
###[Goal]
Rdio was a music streaming service that held apps on several platforms, including iOS and Android phones, Sonos systems and which brand that derived its name from the Japanese word for "six"? You should follow the format instruction in the requestion strictly and wrap the final answer in \boxe... | ###[Observation]
https://en.m.wikipedia.org/wiki/Rdio
URL: https://en.m.wikipedia.org/wiki/Rdio
L1: Document not found | search |
Role: You are an observation generator for a language world model of a web search and page-browsing environment. The action has already occurred; simulate the environment result rather than agent intent.
Input: Use the goal or task context, current observation, action block, and prior turns to reconstruct the current ... | **Turn 1 User**
###[Goal]
In 2018, who was elected mayor of the place Sure of You takes place in? You should follow the format instruction in the requestion strictly and wrap the final answer in \boxed{}.
###[Action]
Use `browser.search` to search the web or retrieval index for relevant documents.
- This advances the t... | ###[Observation]
Doc 1997799
URL: http://womensfictionwriters.com/2016/02/18/guest-post-author-yona-mcdonough-talks-about-choosing-a-setting-for-her-novel/
L1: You know what they say about real estate, right? Location! Location! Location! Well, I think the same can be true for novels. Where we choose to set our novels ... | search |
Role: You are a state-transition world model for an MCP business-tool environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. Apply the recorded a... | **Turn 1 User**
###[Goal]
Register a new pet for existing owner 'Maria Sanchez' named 'Luna' with species 'Canine', breed 'Labrador Retriever', birthdate '2022-03-15', initial weight_history entry of 24.5 kg recorded for today, and status set to 'active'.
###[Action]
Use `list_tools` to list available MCP tools before ... | ###[Observation]
{
"id": 8,
"owner_id": 1,
"name": "Luna",
"species": "Canine",
"breed": "Labrador Retriever",
"birthdate": "2022-03-15T00:00:00",
"status": "active",
"created_at": "2026-08-10T02:37:09.092928",
"updated_at": null
} | mcp |
Role: You are a deterministic language world model for a terminal command-execution environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply the... | **Turn 1 User**
###[Goal]
Please solve this task:
I am a data analyst working with internal communication logs, and I need a high-performance in-memory graph query server. I have two CSV files:
1. `/home/user/nodes.csv` - Contains employee data with headers `id,name,department`.
2. `/home/user/edges.csv` - Contains in... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >... | terminal |
Role: Operate as a language world model for a desktop accessibility-control environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. T... | **Turn 1 User**
###[Observation]
Computer 'Windows Media Player Legacy', b=192,108,1728,972, raw_control='WindowControl'
node=1 generic 'WMPAppHost', raw_control='PaneControl', b=203,153,1717,961
node=2 generic '', raw_control='PaneControl', automation_id='103033244', b=203,207,1717,859
node=3 generic 'LibraryCon... | ###[Observation]
Computer 'Videos Library Locations', b=562,117,1358,890, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=573,162,1347,879
node=2 generic '', raw_control='PaneControl', b=588,357,1202,740
node=3 generic '', raw_control='PaneControl', b=1214,357,1332,395
node=4 generic ... | os |
Role: You are a language world model for a software-engineering editor and shell environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]... | **Turn 1 User**
###[Goal]
<uploaded_files>
/workspace/python__mypy__0.800
</uploaded_files>
I've uploaded a python code repository in the directory python__mypy__0.800. Consider the following PR description:
<pr_description>
Can't assign to ()
Mypy incorrectly complains about assignments to `()`. The simplest example ... | ###[Observation]
EXECUTION RESULT of [execute_bash]:
grep -n "can't assign to ()" /workspace/python__mypy__0.800/mypy/semanal.py
2641: self.fail("can't assign to ()", lval)[Python Interpreter: /opt/miniconda3/envs/testbed/bin/python]
root@openhands-workspace:/workspace/python__mypy__0.800 #
[Command fi... | swe |
Role: You are a language world model for a terminal command-execution environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and any ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a data scientist analyzing spectroscopic signal data. You need to implement a data processing and integration pipeline using multiple languages to extract signal areas and test your numerical methods.
You have been provided with a raw, noisy dataset at `/home/... | ###[Observation]
Smoothed data saved to /home/user/smoothed.csv
Original length: 601, Smoothed length: 597
x,y
-2.98,0.02308651863925248
-2.97,0.015783342865747722
-2.96,0.03296070228732033
-2.95,0.03415773365563824
-2.94,0.014229089001420343
-2.93,0.0219894377139472
-2.92,0.01968681001891574
-2.91,-0.00077514927831228... | terminal |
Role: You are a deterministic language world model for a browser accessibility-control environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply ... | **Turn 1 User**
###[Observation]
Web 'Security & Surveillance - Electronics', focused
[224] banner '', visible
[227] list '', visible
[228] listitem '', visible
[229] link 'My Account', clickable, visible
[230] listitem '', visible
[231] link 'My Wish List 77 items', clickable, visible
StaticText ... | ###[Observation]
Web 'Security & Surveillance - Electronics', focused
[224] banner ''
[227] list ''
[228] listitem ''
[229] link 'My Account', clickable
[230] listitem ''
[231] link 'My Wish List 77 items', clickable
StaticText '77 items'
[233] listitem ''
[234] link 'Sign Out', clickable
... | web |
Role: You are the simulator side of a language world model for a terminal command-execution environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the evidence ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a Data Engineer building an ETL pipeline. Recently, the downstream reporting dashboards have been showing completely blank plots. You've traced the issue to a missing probabilistic data validation step in the ETL pipeline that silently drops all incoming data i... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > [/home/user]$
(exit_code=0) | terminal |
Role: You are a state-transition world model for an MCP business-tool environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. Apply the recorded a... | **Turn 1 User**
###[Goal]
Create or update the tax_profile for employee 'emp-1500' to set jurisdiction_code to 'CA-STATE', filing_status to 'married', and allowances to 3, then recalculate estimated taxes for the next scheduled pay_run this employee is included in.
###[Action]
Use `list_tools` to list available MCP too... | ###[Observation]
{
"employee_id": "emp-1500",
"next_pay_run_id": null,
"estimated_gross_pay": 2769.230769230769,
"estimated_taxes_withheld": 553.8461538461538,
"estimated_net_pay": 2215.3846153846152,
"calculated_at": "2026-07-03T12:48:25.206833"
} | mcp |
Role: You are a deterministic language world model for a browser accessibility-control environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply ... | **Turn 1 User**
###[Observation]
Web 'OpenStreetMap', focused
[44] banner '', visible
[45] heading 'OpenStreetMap logo OpenStreetMap', visible
[46] link 'OpenStreetMap logo OpenStreetMap', clickable, visible
[49] image 'OpenStreetMap logo', visible
StaticText ''
[51] navigation '', visible
[95] link ... | ###[Observation]
Web 'OpenStreetMap', focused
[44] banner '', visible
[45] heading 'OpenStreetMap logo OpenStreetMap', visible
[46] link 'OpenStreetMap logo OpenStreetMap', clickable, visible
[49] image 'OpenStreetMap logo', visible
StaticText ''
[51] navigation '', visible
[95] link 'Edit', clickabl... | web |
Role: Serve as a language world model for an MCP business-tool environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state snapshot. A... | **Turn 1 User**
###[Goal]
Create a new customer profile named 'Acme Retail East' with default pickup address '100 Warehouse Way, Columbus, OH 43215', default delivery address '450 Market St, Philadelphia, PA 19106', preferred carrier 'UPS Ground', and net-30 billing terms, and return the created customer ID.
###[Action... | ###[Observation]
{
"items": [
{
"id": "CARR-UPS",
"name": "UPS",
"created_at": "2024-09-01T08:00:00Z",
"updated_at": null
},
{
"id": "CARR-FDX",
"name": "FedEx",
"created_at": "2024-09-01T08:05:00Z",
"updated_at": null
},
{
"id": "CARR-USPS",
... | mcp |
Role: Act as a language world model for a software-engineering editor and shell environment. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, and p... | **Turn 1 User**
###[Goal]
<uploaded_files>
/workspace/Project-MONAI__MONAI__1.1
</uploaded_files>
I've uploaded a python code repository in the directory Project-MONAI__MONAI__1.1. Consider the following PR description:
<pr_description>
SSIMLoss result does not have gradient
**Describe the bug**
SSIMLoss result does ... | ###[Observation]
EXECUTION RESULT of [str_replace_editor]:
Here's the result of running `cat -n` on /workspace/Project-MONAI__MONAI__1.1/monai/losses/ssim_loss.py:
1 # Copyright (c) MONAI Consortium
2 # Licensed under the Apache License, Version 2.0 (the "License");
3 # you may not use this file except i... | swe |
Role: You are the simulator side of a language world model for a terminal command-execution environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the evidence ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a performance engineer tasked with diagnosing a sudden crash in a numerical simulation pipeline.
The application architecture consists of a frontend ingest service and a backend worker service. Recently, a specific simulation request caused a fatal numerical ... | ###[Observation]
?$Po
;V"r
]D8DQsE
zh:\-
x?Q*
8f>8
UBw" ;
d={}
4`ra
A@f<Q
v11PN
fr~mHTbF=Ow
Vz.%
[fKxt5
z{"req_id": "88B", "weight": 1.797e309, "status": "pending"}
}|M"
o'+h
SEvu
}_4h
_Lj_R6
%WiyM
dZ1j
(exit_code=0) | terminal |
Role: You are a language world model for a terminal command-execution environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and any ... | **Turn 1 User**
###[Goal]
Please solve this task:
I need you to help me clean up a messy bioinformatics dataset before we run our downstream MCMC sampling and regression tests. I'm a bioinformatics analyst, and our pipeline recently suffered from a simulation glitch related to floating-point reduction order, generatin... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > Image created with contaminant sequence: TGACCAATGCGATCG
[/home/user]$
(exit_code=0) | terminal |
Role: You are the simulator side of a language world model for a software-engineering shell testbed. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, the provided `###[Task Description]`, current `###[Action]`, and earlier turns as the ev... | **Turn 1 User**
###[Task Description]
Utility function that verifies division boundaries
Advanced users occasionally set divisions themselves to encode knowledge they have about their data that dask doesn't. However, if these divisions are not set correctly, data loss and therefore incorrect results is a very likely ou... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": " return True\n\n\ndef assert_dask_graph(dask, label):\n if hasattr(dask, \"dask\"):\n dask = dask.dask\n assert isinstance(dask, Mapping)\n for k in dask:\n if isinstance(k, tuple):\n k = k[0]\n if k.startswi... | swe |
Role: You are an observation generator for a language world model of a web search and page-browsing environment. The action has already occurred; simulate the environment result rather than agent intent.
Input: Use the goal or task context, current observation, action block, and prior turns to reconstruct the current ... | **Turn 1 User**
###[Goal]
Find the five tallest residential skyscrapers in Balneário Camboriú, Brazil. For each, provide details about their design, particularly noting if any have a curved façade resulting in apartments of varying sizes, and how many floors each building has.
###[Action]
Use `browser.search` to search... | ###[Observation]
https://www.skyscrapercenter.com/city/balneario-camboriu
URL: https://www.skyscrapercenter.com/city/balneario-camboriu
L1: Document not found | search |
Role: Serve as a language world model for a desktop accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state ... | **Turn 1 User**
###[Observation]
Computer 'Task ID 62 - PowerPoint', b=352,352,1792,1169, raw_control='WindowControl'
node=1 generic 'DropShadowTop', raw_control='PaneControl', b=363,629,1781,638
node=2 generic 'MsoDockTop', raw_control='PaneControl', b=363,363,1781,629
node=3 toolbar '', raw_control='ToolBarContro... | ###[Observation]
Computer 'Context Menu', b=601,639,935,1059, raw_control='MenuControl'
node=1 menu 'Context Menu', raw_control='MenuControl', b=601,639,935,1059
node=2 generic '', raw_control='PaneControl', b=602,640,934,1058
node=3 menu '', raw_control='MenuControl', b=603,641,933,1057
node=4 textbox 'Type ... | os |
Role: You are a deterministic language world model for a desktop accessibility-control environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply ... | **Turn 1 User**
###[Observation]
Computer 'New Tab - Google Chrome', b=0,0,1935,1030, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=9,0,1926,1021
node=2 generic 'New Tab - Google Chrome - Test', raw_control='PaneControl', b=8,0,1927,1022
node=3 generic '', raw_control='PaneControl', b=... | ###[Observation]
Computer '', raw_tag='BODY'
node=1 generic '', raw_tag='DIV', text=' Saved Recents Get app'
node=2 generic 'print', raw_tag='DIV', id='print', class='BPikCc hUbt4d-print'
node=3 generic 'consent-bump', raw_tag='DIV', id='consent-bump', class='Hk4XGb id-consent-bump'
node=4 generic 'Go... | os |
Role: You are a language world model for a terminal command-execution environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and any ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a Site Reliability Engineer (SRE). We have a local development environment where several backend services are supposed to be running. We need to implement a simple uptime monitoring script.
Please write a Python script located at `/home/user/check_uptime.py` t... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > [/home/user]$
(exit_code=0) | terminal |
Role: You are a language world model for a desktop accessibility-control environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and a... | **Turn 1 User**
###[Observation]
Computer 'Run', b=11,777,601,1124, raw_control='WindowControl'
node=1 image '', raw_control='ImageControl', automation_id='12297', b=40,856,88,904
node=2 text 'Type the name of a program, folder, document, or Internet resource, and Windows will open it for you.', raw_control='TextCont... | ###[Observation]
Computer 'Media Player', b=1383,45,1885,358, raw_control='WindowControl'
node=1 window 'Media Player', raw_control='WindowControl', automation_id='TitleBar', b=1730,46,1874,94
node=2 menubar 'System', raw_control='MenuBarControl', automation_id='SystemMenuBar', b=0,0,0,0
node=3 menuitem 'System',... | os |
Role: You are a state-transition world model for a terminal command-execution environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. Apply the re... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a security engineer tasked with rotating credentials and auditing a legacy authentication system. During your review, you discovered that the custom JWT validation logic in `/home/user/service/auth.py` is vulnerable to CWE-347: it improperly verifies cryptograp... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > [/home/user]$ Found compromised token with alg=none
Payload: {'sub': 'hacker1', 'exp': 1700000000}
Found compromised token with... | terminal |
Role: You are an observation generator for a language world model of a mobile accessibility-control environment. The action has already occurred; simulate the environment result rather than agent intent.
Input: Use the goal or task context, current observation, action block, and prior turns to reconstruct the current ... | **Turn 1 User**
###[Observation]
Android '', b=544,584,1059,1508, win=4897, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=544,584,1059,1508, pkg='com.digitalchemy.recorder', actions='4,8,64,16908342', child_ids='1', draw=0, win=4897
node=2 group '', raw_class='FrameLayout', b=544,584,1059,1508, pkg='co... | ###[Observation]
Android '', b=0,0,1080,2400, win=4900, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='com.google.android.gm', actions='4,8,64,16908342', child_ids='1,2,3', draw=0, win=4900
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2337, pkg='com.google.android.gm', actio... | android |
Role: You are a language world model for a browser accessibility-control environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and a... | **Turn 1 User**
###[Observation]
Web 'OpenStreetMap', focused
[44] banner '', visible
[45] heading 'OpenStreetMap logo OpenStreetMap', visible
[46] link 'OpenStreetMap logo OpenStreetMap', clickable, visible
[49] image 'OpenStreetMap logo', visible
StaticText ''
[51] navigation '', visible
[95] link ... | ###[Observation]
Web 'Bushwick, Brooklyn | OpenStreetMap', focused
[44] banner '', visible
[45] heading 'OpenStreetMap logo OpenStreetMap', visible
[46] link 'OpenStreetMap logo OpenStreetMap', clickable, visible
[49] image 'OpenStreetMap logo', visible
StaticText ''
[51] navigation '', visible
[95] ... | web |
Role: You are a language world model for a browser accessibility-control environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and a... | **Turn 1 User**
###[Observation]
Web 'Dashboard / Magento Admin', focused
[148] link 'Magento Admin Panel', clickable, visible
[149] image 'Magento Admin Panel', visible
[150] navigation '', visible
[151] menubar '', visible, orientation='horizontal'
[153] link '\ue604 DASHBOARD', clickable, visible
Static... | ###[Observation]
Web 'Integrations / Extensions / System / Magento Admin', focused
[142] link 'Magento Admin Panel', clickable, visible
[143] image 'Magento Admin Panel', visible
[144] navigation '', visible
[145] menubar '', visible, orientation='horizontal'
[147] link '\ue604 DASHBOARD', clickable, visible
... | web |
Role: You are a deterministic language world model for a software-engineering shell testbed. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply the act... | **Turn 1 User**
###[Task Description]
Recrusive TypedDict crashes Pydantic
### Initial Checks
- [X] I have searched GitHub for a duplicate issue and I'm sure this is something new
- [X] I have searched Google & StackOverflow for a solution and couldn't find anything
- [X] I have read and followed [the docs](https://py... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": "Traceback (most recent call last):\n File \"/tmp/repro.py\", line 2, in <module>\n from pydantic import BaseModel\nModuleNotFoundError: No module named 'pydantic'\n", "stderr": "", "duration_ms": 55} | swe |
Role: Function as the environment model for a browser accessibility environment with reference IDs. Given the action already chosen by another agent, infer the next visible observation without taking agency yourself.
Input: Read the goal or metadata, current observation, action, and history as a single interaction sta... | **Turn 1 User**
###[Observation]
- RootWebArea "Varied Thrush | Sammamish | Thiru Sankaran | Flickr" [ref=e1]
- image [ref=e2]
- navigation "Main" [ref=e3]
- heading "Flickr logo. If you click it, you’ll go home" [ref=e4]
- menubar [ref=e5]
- menuitem "You" [ref=e6]
- menuitem "Explore" [ref=e7]... | ###[Observation]
- RootWebArea "Varied Thrush | Sammamish | Thiru Sankaran | Flickr" [ref=e1]
- image [ref=e2]
- navigation "Main" [ref=e3]
- heading "Flickr logo. If you click it, you’ll go home" [ref=e4]
- menubar [ref=e5]
- menuitem "You" [ref=e6]
- menuitem "Explore" [ref=e7]
- menuite... | web |
Role: You are a state-transition world model for a desktop accessibility-control environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. Apply the... | **Turn 1 User**
###[Observation]
Computer 'Taskbar', b=0,1020,1920,1080, raw_control='PaneControl'
node=1 generic '', raw_control='PaneControl', b=1655,1020,1920,1080
node=2 generic '', raw_control='PaneControl', automation_id='40965', b=599,1020,1100,1080
node=3 generic 'Running applications', raw_control='PaneCon... | ###[Observation]
Computer 'New Tab - Google Chrome', b=0,0,1935,1030, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=9,0,1926,1021
node=2 generic 'New Tab - Google Chrome - Test', raw_control='PaneControl', b=8,0,1927,1022
node=3 generic '', raw_control='PaneControl', b=8,0,1927,1022
... | os |
Role: Operate as a language world model for a web search and page-browsing environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. Th... | **Turn 1 User**
###[Goal]
A Doha-based private satellite media network, financed in part by a ruling monarch of a wealthy peninsula nation and known for reaching hundreds of millions of viewers, received a major international broadcasting award for its coverage of a wave of uprisings that began around 2010 in North Afr... | ###[Observation]
https://foreignpolicy.com/2011/01/06/obamas-arab-spring/
URL: https://foreignpolicy.com/2011/01/06/obamas-arab-spring/
L1: Document not found | search |
Role: You are a deterministic language world model for a terminal command-execution environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply the... | **Turn 1 User**
###[Goal]
Please solve this task:
You are an engineer setting up a polyglot build system from scratch for a high-performance mathematical pipeline. We have a mathematical workflow that requires chaining operations written in Go (for concurrency) and Rust (for memory-safe processing), orchestrated by Py... | ###[Observation]
#[no_mangle]
pub extern "C" fn compute_gcds(arr: *mut u32, length: usize, val: u32) {
let slice = unsafe { std::slice::from_raw_parts_mut(arr, length) };
for i in 0..length {
let current = &mut slice[i];
let mut a = *current;
let mut b = val;
let temp_ref = cur... | terminal |
Role: You are a deterministic language world model for a software-engineering shell testbed. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply the act... | **Turn 1 User**
###[Task Description]
Need to deepcopy data in RandCropByPosNegLabeld transform
**Describe the bug**
Currently, when we crop samples in `RandCropByPosNegLabeld` transform, it uses shallow copy for meta_data and other information, then if we change some value in a sample, all the samples are changed.
N... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": "485:class RandSpatialCropSamplesd(Randomizable, MapTransform):\n486- \"\"\"\n487- Dictionary-based version :py:class:`monai.transforms.RandSpatialCropSamples`.\n488- Crop image with random size or specific size ROI to generate a list of N samp... | swe |
Role: Serve as a language world model for an MCP business-tool environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state snapshot. A... | **Turn 1 User**
###[Goal]
Update my customer profile preferences to set 'preferred_staff' to 'Michael Lopez', note a sensitivity to 'lavender essential oils', and store that I prefer 'low music volume' in treatment rooms.
###[Action]
Use `list_tools` to list available MCP tools before selecting a specific tool.
- This ... | ###[Observation]
{
"customer_preferences": {
"id": 1,
"customer_id": 1,
"preferred_staff_member_id": 2,
"notes": "Prefers low music volume in treatment rooms",
"scent_sensitivities": "lavender essential oils",
"music_volume_preference": "low",
"created_at": "2025-12-01T13:00:00",
"upda... | mcp |
Role: You are a post-action environment simulator for an MCP business-tool environment. The input action is already fixed; your role is to predict what the environment exposes next.
Input: Use the available goal, metadata, latest `###[Observation]`, action text, and history as state evidence. Do not add, skip, reorder... | **Turn 1 User**
###[Goal]
Record completion of work order ID 'SH-307-WO-1145' by setting status to 'Completed' on 2025-01-11, adding technician note 'Replaced P-trap and tightened fittings; no further leaks observed', and attach a vendor invoice for $185 coded to GL account 'Maintenance – Plumbing'.
###[Action]
Use `li... | ###[Observation]
{
"id": 4,
"vendor_id": 1,
"property_id": 4,
"invoice_number": null,
"invoice_date": "2025-01-11T00:00:00",
"amount": 185.0,
"gl_account_id": 2,
"description": "Work Order SH-307-WO-1145: Kitchen sink leak repair – Replaced P-trap and tightened fittings",
"created_at": "2026-07-03T10:... | mcp |
Role: You are a state-transition world model for a mobile accessibility-control environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. Apply the ... | **Turn 1 User**
###[Observation]
Android '', b=0,0,1080,2400, win=1997, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='com.google.android.apps.nexuslauncher', actions='4,8,64,16908342', child_ids='1', draw=0, win=1997
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2400, pkg='c... | ###[Observation]
Android '', b=0,0,1080,2400, win=2107, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='com.plant.identify.care.app', actions='4,8,64,16908342', child_ids='1,2', draw=0, win=2107
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2337, pkg='com.plant.identify.care.a... | android |
Role: You are a deterministic language world model for a browser accessibility-control environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply ... | **Turn 1 User**
###[Observation]
Web 'Dashboard / Magento Admin', focused
[148] link 'Magento Admin Panel', clickable, visible
[149] image 'Magento Admin Panel', visible
[150] navigation '', visible
[151] menubar '', visible, orientation='horizontal'
[153] link '\ue604 DASHBOARD', clickable, visible
Static... | ###[Observation]
Web 'Bestsellers Report / Products / Reports / Magento Admin', focused
[149] link 'Magento Admin Panel', clickable, visible
[150] image 'Magento Admin Panel', visible
[151] navigation '', visible
[152] menubar '', visible, orientation='horizontal'
[154] link '\ue604 DASHBOARD', clickable, visi... | web |
Role: You are the simulator side of a language world model for a software-engineering shell testbed. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, the provided `###[Task Description]`, current `###[Action]`, and earlier turns as the ev... | **Turn 1 User**
###[Task Description]
CopyItemsd transform saves data into same key
**Describe the bug**
This line of `CopyItemsd` seems not correct:
https://github.com/Project-MONAI/MONAI/blob/dev/monai/transforms/utility/dictionary.py#L740
It saves all the data of keys into the same `new_key`.
###[Action]
Use `... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": "class CopyItemsd(MapTransform):\n \"\"\"\n Copy specified items from data dictionary and save with different key names.\n It can copy several items together and copy several times.\n\n \"\"\"\n\n def __init__(\n self, keys: KeysCo... | swe |
Role: You are a deterministic language world model for a web search and page-browsing environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply t... | **Turn 1 User**
###[Goal]
What is considered a greater veterinary problem among companion animals than nutritional deficiency? You should follow the format instruction in the requestion strictly and wrap the final answer in \boxed{}.
###[Action]
Use `browser.search` to search the web or retrieval index for relevant doc... | ###[Observation]
Doc 4967208
URL: https://vcahospitals.com/plainfield/services/pet-counseling
L1: Our hospital offers counseling for issues that may be affecting your pet. These issues could range from a pet being overweight and needing nutritional counseling to changes in behavior that require behavior modification. I... | search |
Role: You are a state-transition world model for a software-engineering editor and shell environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. A... | **Turn 1 User**
###[Goal]
<uploaded_files>
/workspace/Project-MONAI__MONAI__1.0
</uploaded_files>
I've uploaded a python code repository in the directory Project-MONAI__MONAI__1.0. Consider the following PR description:
<pr_description>
ContrastiveLoss with dynamic batchsize
**Is your feature request related to a prob... | ###[Observation]
EXECUTION RESULT of [str_replace_editor]:
[Code executed successfully with no output] | swe |
Role: You are a deterministic language world model for a software-engineering shell testbed. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply the act... | **Turn 1 User**
###[Task Description]
BUG: "ValueError: The truth value ... is ambiguous" when using .loc setitem
### Pandas version checks
- [X] I have checked that this issue has not already been reported.
- [X] I have confirmed this bug exists on the [latest version](https://pandas.pydata.org/docs/whatsnew/ind... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": " Insert new row(s) or column(s) into the Series or DataFrame.\n \"\"\"\n from pandas import Series\n\n # reindex the axis to the new value\n # and set inplace\n if self.ndim == 1:\n index = self.obj.... | swe |
Role: You are a post-action environment simulator for a terminal command-execution environment. The input action is already fixed; your role is to predict what the environment exposes next.
Input: Use the available goal, metadata, latest `###[Observation]`, action text, and history as state evidence. Do not add, skip,... | **Turn 1 User**
###[Goal]
Please solve this task:
You are an open-source maintainer reviewing a pull request for a multi-language microservices repository. The contributor has modified some gRPC protobuf definitions, but the CI pipeline is failing.
The CI pipeline relies on a bash script, `/home/user/ci_check.sh`, wh... | ###[Observation]
syntax = "proto3";
package auth;
service Auth {
rpc Login (LoginRequest) returns (LoginResponse);
rpc Logout (LogoutRequest) returns (LogoutResponse);
}
(exit_code=0) | terminal |
Role: Act as a language world model for a terminal command-execution environment. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, and prior observ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a platform engineer responsible for maintaining our CI/CD pipelines. We are building a new artifact validation system.
There are two components to this system:
1. A small Rust CLI tool used by build agents to compute Adler-32 checksums of artifacts.
2. A Pyth... | ###[Observation]
syntax = "proto3";
package artifact;
service ArtifactValidator {
rpc Validate (ValidationRequest) returns (ValidationResponse);
}
message ValidationRequest {
string client_id = 1;
string filename = 2;
string checksum = 3;
}
message ValidationResponse {
bool valid = 1;
}
(exit_code=0) | terminal |
Role: You are a language world model for an MCP business-tool environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and any earlier ... | **Turn 1 User**
###[Goal]
Buy exactly $150 worth of Bitcoin (BTC) at the current market price using my linked bank account and leave the purchased BTC in my Coinbase wallet.
###[Action]
Use `list_tools` to list available MCP tools before selecting a specific tool.
- This advances the task or changes state.
- Call:
```t... | ###[Observation]
{
"payment_methods": [
{
"id": 1,
"type": "bank_account",
"provider": "Chase Bank",
"last4": "4321",
"is_default": false,
"status": "active",
"created_at": "2026-04-04T09:29:57",
"updated_at": "2026-06-23T09:29:57"
},
{
"id": 2,
... | mcp |
Role: Act as a language world model for a desktop accessibility-control environment. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, and prior obs... | **Turn 1 User**
###[Observation]
Computer 'Task ID - Protected View - Excel', b=-11,-11,1931,1019, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=0,975,1920,1008
node=2 toolbar '', raw_control='ToolBarControl', b=0,975,1920,1008
node=3 generic 'Status Bar', raw_control='PaneControl', ... | ###[Observation]
Computer 'Task ID - Excel', b=-11,-11,1931,1019, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=0,975,1920,1008
node=2 toolbar '', raw_control='ToolBarControl', b=0,975,1920,1008
node=3 generic 'Status Bar', raw_control='PaneControl', b=0,975,1920,1008
node=4 gene... | os |
Role: Act as a language world model for a mobile accessibility-control environment. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, and prior obse... | **Turn 1 User**
###[Observation]
Android '', b=0,0,1080,2400, win=328, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='com.google.android.calendar', actions='4,8,64,16908342', child_ids='1', draw=0, win=328
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2400, pkg='com.google.an... | ###[Observation]
Android '', b=28,597,1052,1857, win=332, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=28,597,1052,1857, pkg='com.google.android.calendar', actions='4,8,64,16908342', child_ids='1', draw=0, win=332
node=2 group '', raw_class='LinearLayout', b=91,807,989,1647, pkg='com.google.android.ca... | android |
Role: Function as the environment model for a software-engineering shell testbed. Given the action already chosen by another agent, infer the next visible observation without taking agency yourself.
Input: Read the goal or metadata, current observation, action, and history as a single interaction state. Only the suppl... | **Turn 1 User**
###[Task Description]
[Bug] AssertionError triggered with multiple configs from a Config Group specified at CLI
# 🐛 Bug
## Description
An assertion error is raised if the CLI is used to select multiple options from a Config Group:
```python
$ cat my_app.py
import hydra
from omegaconf import Ome... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": "\n override: bool = False\n deleted: Optional[bool] = None\n\n config_name_overridden: bool = field(default=False, compare=False, repr=False)\n\n def __post_init__(self) -> None:\n assert self.group is not None and self.group != \"\"... | swe |
Role: Operate as a language world model for a browser accessibility-control environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. T... | **Turn 1 User**
###[Observation]
Web '#000000299 / Orders / Operations / Sales / Magento Admin', focused
[186] link 'Magento Admin Panel', clickable, visible
[187] image 'Magento Admin Panel', visible
[188] navigation '', visible
[189] menubar '', visible, orientation='horizontal'
[191] link '\ue604 DASHBOARD'... | ###[Observation]
Web '#000000299 / Orders / Operations / Sales / Magento Admin', focused
[186] link 'Magento Admin Panel', clickable, visible
[187] image 'Magento Admin Panel', visible
[188] navigation '', visible
[189] menubar '', visible, orientation='horizontal'
[191] link '\ue604 DASHBOARD', clickable, vis... | web |
Role: You are a deterministic language world model for a web search and page-browsing environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply t... | **Turn 1 User**
###[Goal]
When did the person who plays the receptionist in How I Met Your Mother release her first album? You should follow the format instruction in the requestion strictly and wrap the final answer in \boxed{}.
###[Action]
Use `browser.search` to search the web or retrieval index for relevant documen... | ###[Observation]
https://www.britannica.com/topic/Britney-Spears
URL: https://www.britannica.com/topic/Britney-Spears
L1: Document not found | search |
Role: You are a language world model for a software-engineering editor and shell environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]... | **Turn 1 User**
###[Goal]
<uploaded_files>
/workspace/getmoto__moto__3.1
</uploaded_files>
I've uploaded a python code repository in the directory getmoto__moto__3.1. Consider the following PR description:
<pr_description>
KeyError: 'clustername' (instead of `DBClusterNotFoundFault`) thrown when calling `describe_db_c... | ###[Observation]
EXECUTION RESULT of [execute_bash]:
grep -nr 'def describe_db_clusters' /workspace/getmoto__moto__3.1/moto
/workspace/getmoto__moto__3.1/moto/rds/models.py:1797: def describe_db_clusters(self, cluster_identifier):
/workspace/getmoto__moto__3.1/moto/rds/responses.py:511: def describe_db_clusters... | swe |
Role: You are a state-transition world model for a terminal command-execution environment. Your prediction is the next observation after the given action, not a recommendation or solution.
Input: Treat the goal, metadata, latest observation, action text, and earlier turns as the current state description. Apply the re... | **Turn 1 User**
###[Goal]
Please solve this task:
You are helping a researcher organize and process a multimodal dataset consisting of sensor logs and an audio recording.
The dataset contains three files:
1. `/app/input.wav`: An audio recording from the experiment.
2. `/app/events.csv`: Sensor readings. It has colum... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > [/home/user]$ Created WAV: 154350 frames, 44100 Hz, duration = 3.5
Created events.csv
Created labels.csv
Active sensor sum: 45.5
Duration: 3.5
Expected result: 159.25
... | terminal |
Role: You are an observation generator for a language world model of a terminal command-execution environment. The action has already occurred; simulate the environment result rather than agent intent.
Input: Use the goal or task context, current observation, action block, and prior turns to reconstruct the current st... | **Turn 1 User**
###[Goal]
Please solve this task:
You are acting as a capacity planner analyzing resource usage across different geographical regions. To securely collect and route these metrics, you need to set up a secure aggregation endpoint.
Your task is to restore a configuration backup, write a Go-based secure ... | ###[Observation]
> > > > > > > > > > > > > > > >
.bash_logout .profile config.json
.bashrc certs/ metrics_backup.tar.gz
> > > > >
.bash_logout .profile config.json
.bashrc certs/ metrics_backup.tar.gz
> > > >... | terminal |
Role: Operate as a language world model for a terminal command-execution environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. The ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a machine learning engineer preparing a training dataset for a surrogate model. You have been given a large dataset of noisy time-series signals, and you need to extract the underlying physical parameters for each signal to serve as the ground-truth labels for ... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > [/home/user]$ Testing on first 5 signals:
idx | a_fit | b_fit | omega_fit | SSE
0 | 2.1484 | 0.4432 | 4.3645 | 8.069813
Reconstructio... | terminal |
Role: You are a deterministic language world model for a software-engineering editor and shell environment. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record... | **Turn 1 User**
###[Goal]
<uploaded_files>
/workspace/iterative__dvc__2.8
</uploaded_files>
I've uploaded a python code repository in the directory iterative__dvc__2.8. Consider the following PR description:
<pr_description>
Negative numbers from a Python file are not recognized as parameters
# Bug Report
## dvc re... | ###[Observation]
EXECUTION RESULT of [str_replace_editor]:
ERROR:
The `view_range` parameter is not allowed when `path` points to a directory. | swe |
Role: You are a language world model for a web search and page-browsing environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and an... | **Turn 1 User**
###[Goal]
Find universities in the UK named after knighted British eye doctors or ophthalmologists. Then check if any Essex-based institution merged with this university in a year spanning February 2024.
###[Action]
First, use `browser.search` to search the web or retrieval index for relevant documents.... | ###[Observation]
Doc 6228972
URL: https://www.issoct.com/biographies/pearse-keane-md
L1: Mr. Pearse Keane, MD, MSc, FRCOphth, MRCSI
L2: Consultant Ophthalmologist/NHR Clinician Scientist, Moorfields Eye Hospital/Institute of Ophthalmology University College
L3: Mr Keane is an NIHR Clinician Scientist and Honorary Consu... | search |
Role: Serve as a language world model for a desktop accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state ... | **Turn 1 User**
###[Observation]
Computer 'Computer-Use-Agents - Visual Studio Code [Administrator]', b=450,107,1452,895, raw_control='PaneControl'
node=1 document 'Computer-Use-Agents - Visual Studio Code [Administrator]', raw_control='DocumentControl', automation_id='23466992', b=450,107,1452,896
node=2 group '', ... | ###[Observation]
Computer 'Taskbar', b=0,1128,1920,1200, raw_control='PaneControl'
node=1 generic '', raw_control='PaneControl', b=1483,1128,1920,1200
node=2 generic '', raw_control='PaneControl', automation_id='40965', b=593,1128,1128,1200
node=3 generic 'Running applications', raw_control='PaneControl', b=593,112... | os |
Role: You are a language world model for a terminal command-execution environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and any ... | **Turn 1 User**
###[Goal]
Please solve this task:
I am researching thermal diffusion across molecular network structures. I have an HDF5 file containing the adjacency matrix of a network and its initial temperature distribution, but my current explicit simulation scheme keeps blowing up for certain configurations due ... | ###[Observation]
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > [/home/user]$ Network size: 5 nodes
Adjacency matrix:
[[0. 1. 1. 1. 1.]
[1. 0. 0. 0. 0.]
[1. 0. 0. 0. 0.]
[1. 0. 0. ... | terminal |
Role: You are a language world model for a web search and page-browsing environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and an... | **Turn 1 User**
###[Goal]
Find universities in the Mid-Eastern Athletic Conference (MEAC) that are located in cities with the same name as their state, and identify which one upset a #2 seed in its first NCAA tournament appearance in the 2010s.
###[Action]
Use `browser.search` to search the web or retrieval index for r... | ###[Observation]
https://meacsports.com/sports/2024/6/5/members.aspx
URL: https://meacsports.com/sports/2024/6/5/members.aspx
L1: Document not found | search |
Role: Serve as a language world model for a browser accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state ... | **Turn 1 User**
###[Observation]
Web 'Projects · Dashboard · GitLab', focused
[51] banner '', visible
[52] link 'Skip to content', clickable
StaticText 'GitLab'
[58] link 'Dashboard', clickable, visible
[59] image '', visible
[62] list '', visible
[63] listitem '', visible
[64] button '', clickable, ... | ###[Observation]
Web 'Issues · Oliver Jones / timeit · GitLab', focused
[64] banner '', visible
[65] link 'Skip to content', clickable
StaticText 'GitLab'
[71] link 'Dashboard', clickable, visible
[72] image '', visible
[75] list '', visible
[76] listitem '', visible
[77] button '', clickable, visibl... | web |
Role: Operate as a language world model for a desktop accessibility-control environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. T... | **Turn 1 User**
###[Observation]
Computer 'Save As', b=0,0,960,760, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=11,119,949,650
node=2 generic 'Explorer Pane', raw_control='PaneControl', automation_id='main', b=11,119,949,650
node=3 generic '', raw_control='PaneControl', b=199,478,9... | ###[Observation]
Computer 'Task ID 49 - AutoRecovered - Word', b=-11,-11,1931,1139, raw_control='WindowControl'
node=1 generic 'DropShadowTop', raw_control='PaneControl', b=0,359,1920,368
node=2 generic 'MsoDockLeft', raw_control='PaneControl', b=0,359,525,1095
node=3 toolbar '', raw_control='ToolBarControl', b=0,3... | os |
Role: Operate as a language world model for a browser accessibility-control environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. T... | **Turn 1 User**
###[Observation]
Web 'Bestsellers Report / Products / Reports / Magento Admin', focused
[149] link 'Magento Admin Panel', clickable, visible
[150] image 'Magento Admin Panel', visible
[151] navigation '', visible
[152] menubar '', visible, orientation='horizontal'
[154] link '\ue604 DASHBOARD',... | ###[Observation]
Web 'Bestsellers Report / Products / Reports / Magento Admin', focused
[149] link 'Magento Admin Panel', clickable, visible
[150] image 'Magento Admin Panel', visible
[151] navigation '', visible
[152] menubar '', visible, orientation='horizontal'
[154] link '\ue604 DASHBOARD', clickable, visi... | web |
Role: Operate as a language world model for a desktop accessibility-control environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. T... | **Turn 1 User**
###[Observation]
Computer 'Taskbar', b=0,1020,1920,1080, raw_control='PaneControl'
node=1 generic '', raw_control='PaneControl', b=1655,1020,1920,1080
node=2 generic '', raw_control='PaneControl', automation_id='40965', b=599,1020,1100,1080
node=3 generic 'Running applications', raw_control='PaneCon... | ###[Observation]
Computer 'Home - File Explorer', b=588,360,1902,1020, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=958,360,1893,403
node=2 generic 'Home', raw_control='PaneControl', b=597,530,1893,1011
node=3 generic '', raw_control='PaneControl', b=597,530,1893,1011
node=4 generi... | os |
Role: You are an observation generator for a language world model of a mobile accessibility-control environment. The action has already occurred; simulate the environment result rather than agent intent.
Input: Use the goal or task context, current observation, action block, and prior turns to reconstruct the current ... | **Turn 1 User**
###[Observation]
Android '', b=0,0,1080,2400, win=397, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='android', actions='4,8,64,16908342', child_ids='1', draw=0, win=397
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2400, pkg='android', actions='4,8,64,1690834... | ###[Observation]
Android '', b=0,0,1080,2400, win=399, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='com.google.android.gm', actions='4,8,64,16908342', child_ids='1,2,4', draw=0, win=399
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2337, pkg='com.google.android.gm', actions... | android |
Role: Function as the environment model for a terminal command-execution environment. Given the action already chosen by another agent, infer the next visible observation without taking agency yourself.
Input: Read the goal or metadata, current observation, action, and history as a single interaction state. Only the s... | **Turn 1 User**
###[Goal]
Please solve this task:
You are helping a researcher organize and validate a dataset. The researcher wrote a Python script `/home/user/analyze.py` to analyze `/home/user/dataset.csv`. The script uses bootstrap sampling to find the 95% confidence interval (using the 2.5th and 97.5th percentile... | ###[Observation]
(no output)
(exit_code=0) | terminal |
Role: You are a post-action environment simulator for an MCP business-tool environment. The input action is already fixed; your role is to predict what the environment exposes next.
Input: Use the available goal, metadata, latest `###[Observation]`, action text, and history as state evidence. Do not add, skip, reorder... | **Turn 1 User**
###[Goal]
Create a workflow called 'High-Value Deal Alert' that triggers when my 'Salesforce' connector emits an event 'opportunity_updated', adds a condition step that checks if opportunity.amount is greater than 50000 and opportunity.stage equals 'Negotiation', and if true executes a 'Slack' connector... | ###[Observation]
{
"workflows": [
{
"id": 1,
"workspace_id": 1,
"name": "Website Lead to CRM and Slack",
"description": "Webhook to create HubSpot contact and notify Slack",
"is_active": true,
"created_by_user_id": 1,
"created_at": "2024-12-15T09:10:00",
"updated_at... | mcp |
Role: Serve as a language world model for a desktop accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state ... | **Turn 1 User**
###[Observation]
Computer 'Ribbon Display Options', b=1612,261,1916,490, raw_control='MenuControl'
###[Action]
First, use `mouse_move` to move the mouse pointer by a delta.
- This advances the task or changes state.
- Args: `delta`=[4,2].
- Call:
```text
mouse_move(delta=[4,2])
```
Finally, use `mouse_... | ###[Observation]
Computer 'Slide Size', b=1426,237,1672,432, raw_control='MenuControl' | os |
Role: You are a deterministic language world model for a software-engineering shell testbed. Predict the environment observation caused by the already-issued action, and nothing else.
Input: Use all provided goal or task context, recent observation, current action, and previous turns as the state record. Apply the act... | **Turn 1 User**
###[Task Description]
Aurora Postgres Does Not Support Mock ServerlessV2
From what I can tell, Aurora Postgres does not support ServerlessV2ScalingConfiguration like Neptune does. Support for this would be great!
###[Action]
First, use `shell_exec` to execute a shell command in the SWE task environment... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": "tests/test_cloudformation/fixtures/rds_mysql_with_db_parameter_group.py\ntests/test_cloudformation/fixtures/rds_mysql_with_read_replica.py\ntests/test_cloudwatch/test_cloudwatch_dashboards.py\ntests/test_neptune/__init__.py\ntests/test_neptune/test_clu... | swe |
Role: Serve as a language world model for a browser accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state ... | **Turn 1 User**
###[Observation]
Web 'New Customer / Customers / Customers / Magento Admin', focused
[176] link 'Magento Admin Panel', clickable, visible
[177] image 'Magento Admin Panel', visible
[178] navigation '', visible
[179] menubar '', visible, orientation='horizontal'
[181] link '\ue604 DASHBOARD', cl... | ###[Observation]
Web 'New Customer / Customers / Customers / Magento Admin', focused
[176] link 'Magento Admin Panel', clickable, visible
[177] image 'Magento Admin Panel', visible
[178] navigation '', visible
[179] menubar '', visible, orientation='horizontal'
[181] link '\ue604 DASHBOARD', clickable, visible... | web |
Role: You are a post-action environment simulator for an MCP business-tool environment. The input action is already fixed; your role is to predict what the environment exposes next.
Input: Use the available goal, metadata, latest `###[Observation]`, action text, and history as state evidence. Do not add, skip, reorder... | **Turn 1 User**
###[Goal]
Create a workflow called 'High-Value Deal Alert' that triggers when my 'Salesforce' connector emits an event 'opportunity_updated', adds a condition step that checks if opportunity.amount is greater than 50000 and opportunity.stage equals 'Negotiation', and if true executes a 'Slack' connector... | ###[Observation]
{
"operations": [
{
"id": 6,
"connector_id": 5,
"name": "opportunity_updated",
"type": "event",
"description": "Salesforce opportunity updated event",
"config_schema": {},
"input_schema": {
"type": "object",
"properties": {
"oppo... | mcp |
Role: You are the simulator side of a language world model for a browser accessibility-control environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the eviden... | **Turn 1 User**
###[Observation]
Web 'User:The other Kiwix guy/Landing', focused
[30] Section '', visible
[32] LabelText '', clickable, visible
StaticText '🔍'
[33] textbox "Search 'Wikipedia'", clickable, visible
[38] link 'Go to welcome page', clickable, visible
[39] button '🏠', visible
[40] link "Go to ... | ###[Observation]
Web 'Snatch (weightlifting)', focused
[32] Section '', visible
[34] LabelText '', clickable, visible
StaticText '🔍'
[35] textbox "Search 'Wikipedia'", clickable, visible
[40] link 'Go to welcome page', clickable, visible
[41] button '🏠', visible
[42] link "Go to the main page of 'Wikipedi... | web |
Role: You are an observation generator for a language world model of a software-engineering shell testbed. The action has already occurred; simulate the environment result rather than agent intent.
Input: Use the goal or task context, current observation, action block, and prior turns to reconstruct the current state.... | **Turn 1 User**
###[Task Description]
IoT - thingId not present after create_thing()
```
import boto3
from moto import mock_iot
@mock_iot
def test_create_thing():
client = boto3.client("iot", region_name="ap-northeast-1")
thing = client.create_thing(
thingName="IDontHaveAThingId"
)
... | ###[Observation]
{"status": "SUCCEEDED", "exit_code": 0, "stdout": "import boto3\n\nfrom moto import mock_iot\nfrom moto.core import DEFAULT_ACCOUNT_ID as ACCOUNT_ID\n\n\n@mock_iot\ndef test_create_thing():\n client = boto3.client(\"iot\", region_name=\"ap-northeast-1\")\n name = \"my-thing\"\n\n thing = clien... | swe |
Role: You are the simulator side of a language world model for a desktop accessibility-control environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the eviden... | **Turn 1 User**
###[Observation]
Computer 'Snipping Tool Overlay', b=0,0,1920,1080, raw_control='WindowControl'
node=1 generic 'DesktopWindowXamlSource', raw_control='PaneControl', b=0,0,1920,1080
node=2 generic '', raw_control='PaneControl', b=0,0,1280,720
node=3 image '', raw_control='ImageControl', b=0,0,1920,... | ###[Observation]
Computer 'Taskbar', b=0,1008,1920,1080, raw_control='PaneControl'
node=1 generic 'PopupHost', raw_control='PaneControl', b=0,1008,1280,1056
node=2 window 'Popup', raw_control='WindowControl', b=0,0,0,0
node=3 tooltip '', raw_control='ToolTipControl', b=915,942,1046,991
node=4 text 'File Explo... | os |
Role: You are the simulator side of a language world model for a browser accessibility-control environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the eviden... | **Turn 1 User**
###[Observation]
Web 'One Stop Market', focused
[222] banner '', visible
[225] list '', visible
[226] listitem '', visible
[227] link 'My Account', clickable, visible
[228] listitem '', visible
[229] link 'My Wish List 37 items', clickable, visible
StaticText '37 items'
[231] li... | ###[Observation]
Web 'Clothing, Shoes & Jewelry', focused
[224] banner '', visible
[227] list '', visible
[228] listitem '', visible
[229] link 'My Account', clickable, visible
[230] listitem '', visible
[231] link 'My Wish List 37 items', clickable, visible
StaticText '37 items'
[233] listitem... | web |
Role: You are a language world model for a desktop accessibility-control environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and a... | **Turn 1 User**
###[Observation]
Computer 'Task ID 41 - PowerPoint', b=-11,-11,1931,1019, raw_control='WindowControl'
node=1 window 'Tell your story with video', raw_control='WindowControl', b=1269,112,1711,640
node=2 window 'Tell your story with video', raw_control='WindowControl', b=1269,112,1711,640
node=3 gro... | ###[Observation]
Computer '', b=30,134,774,821, raw_control='MenuControl' | os |
Role: You are a language world model for an MCP business-tool environment. Your responsibility is to simulate the environment after an action has already been issued, not to act as the agent.
Input: This prompt accepts the goal or task context, the latest `###[Observation]`, the current `###[Action]`, and any earlier ... | **Turn 1 User**
###[Goal]
Buy exactly $150 worth of Bitcoin (BTC) at the current market price using my linked bank account and leave the purchased BTC in my Coinbase wallet.
###[Action]
Use `list_tools` to list available MCP tools before selecting a specific tool.
- This advances the task or changes state.
- Call:
```t... | ###[Observation]
{
"id": 2,
"type": "card",
"provider": "Visa",
"last4": "1111",
"is_default": true,
"status": "active",
"created_at": "2026-05-04T09:29:57",
"updated_at": "2026-06-28T09:29:57"
} | mcp |
Role: You are the simulator side of a language world model for a desktop accessibility-control environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the eviden... | **Turn 1 User**
###[Observation]
Computer 'Task ID 72 - Word', b=-9,-9,1929,1029, raw_control='WindowControl'
node=1 generic 'DropShadowTop', raw_control='PaneControl', b=0,260,1920,268
node=2 generic 'MsoDockTop', raw_control='PaneControl', b=0,0,1920,260
node=3 toolbar '', raw_control='ToolBarControl', b=0,222,19... | ###[Observation]
Computer '', b=0,0,0,0, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=0,0,0,0 | os |
Role: Serve as a language world model for a mobile accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state s... | **Turn 1 User**
###[Observation]
Android '', b=0,0,1080,2400, win=1104, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='com.babbel.mobile.android.en', actions='4,8,64,16908342', child_ids='1,2,3', draw=0, win=1104
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2337, pkg='com.ba... | ###[Observation]
Android '', b=0,0,1080,2400, win=1103, layer=0, type=1
node=1 group '', raw_class='FrameLayout', b=0,0,1080,2400, pkg='com.babbel.mobile.android.en', actions='4,8,64,16908342', child_ids='1,2,4', draw=0, win=1103
node=2 group '', raw_class='LinearLayout', b=0,0,1080,2337, pkg='com.babbel.mobile.andr... | android |
Role: Operate as a language world model for a browser accessibility-control environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. T... | **Turn 1 User**
###[Observation]
Web 'Edit Profile · User Settings · GitLab', focused
[59] banner '', visible
[60] link 'Skip to content', clickable
StaticText 'GitLab'
[66] link 'Dashboard', clickable, visible
[67] image '', visible
[70] list '', visible
[71] listitem '', visible
[72] button '', cli... | ###[Observation]
Web 'Edit Profile · User Settings · GitLab', focused
[59] banner '', visible
[60] link 'Skip to content', clickable
StaticText 'GitLab'
[66] link 'Dashboard', clickable, visible
[67] image '', visible
[70] list '', visible
[71] listitem '', visible
[72] button '', clickable, visible,... | web |
Role: Operate as a language world model for a desktop accessibility-control environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. T... | **Turn 1 User**
###[Observation]
Computer 'Task ID 63 - Excel', b=-11,-11,1931,1139, raw_control='WindowControl'
node=1 generic 'DropShadowTop', raw_control='PaneControl', b=0,267,1920,276
node=2 generic '', raw_control='PaneControl', b=0,324,400,1095
node=3 toolbar '', raw_control='ToolBarControl', b=0,324,400,109... | ###[Observation]
Computer '&Insert Pie or Doughnut Chart', b=761,237,1017,711, raw_control='MenuControl' | os |
Role: Function as the environment model for a web search and page-browsing environment. Given the action already chosen by another agent, infer the next visible observation without taking agency yourself.
Input: Read the goal or metadata, current observation, action, and history as a single interaction state. Only the... | **Turn 1 User**
###[Goal]
What Alabama governor ran against Fob James in 1998? You should follow the format instruction in the requestion strictly and wrap the final answer in \boxed{}.
###[Action]
Use `browser.search` to search the web or retrieval index for relevant documents.
- This advances the task or changes stat... | ###[Observation]
Doc 5689
URL: https://en.wikipedia.org/wiki/Don_Siegelman
L1: Don Siegelman
L2:
L3: 51st Governor of Alabama
L4: In office
L5: January 18, 1999 – January 20, 2003
L6: Lieutenant Steve Windom
L7: Preceded by Fob James
L8: Succeeded by Bob Riley
L9: 26th Lieutenant Governor of Alaba... | search |
Role: You are an observation generator for a language world model of an MCP business-tool environment. The action has already occurred; simulate the environment result rather than agent intent.
Input: Use the goal or task context, current observation, action block, and prior turns to reconstruct the current state. App... | **Turn 1 User**
###[Goal]
As a student, retrieve all upcoming confirmed bookings for the next 14 days and cancel the one scheduled for '2025-01-15T16:00:00' with tutor 'Sarah Kim', specifying the cancellation reason as 'school exam rescheduled'.
###[Action]
Use `list_tools` to list available MCP tools before selecting ... | ###[Observation]
{
"bookings": []
} | mcp |
Role: Operate as a language world model for an MCP business-tool environment. You model how the environment changes after an issued action and report only the resulting observation.
Input: Interpret the goal or task metadata, latest observation, action block, and earlier turns as a persistent state trace. The current ... | **Turn 1 User**
###[Goal]
Create a new subscription for customer 'Acme Design Studio' using the 'Content Marketing Retainer' monthly price_plan with start_date '2026-01-01', trial_end '2026-01-15', proration_policy 'full', and automatic renewal each month on the same day, then schedule automatic recurring invoices for ... | ###[Observation]
{
"items": [
{
"id": 1,
"product_id": 1,
"name": "Content Marketing Retainer - Monthly",
"billing_cycle": "monthly",
"unit_price": 2500.0,
"currency": "USD",
"tax_category": "standard",
"is_recurring": true,
"created_at": "2024-12-20T10:10:00"... | mcp |
Role: Act as a language world model for a terminal command-execution environment. You predict the post-action observation of the environment; you are not the acting agent, planner, or tool executor.
Input: Use the goal or task description, the most recent `###[Observation]`, the current `###[Action]`, and prior observ... | **Turn 1 User**
###[Goal]
Please solve this task:
You are a data scientist building a reproducible computational pipeline for conformational sampling of protein structures. Our legacy system relies on an opaque, stripped binary `/app/mcmc_evaluator` which runs a Parallel Tempering MCMC sampling simulation over sequenc... | ###[Observation]
/usr/bin/python3
Python 3.10.12
---
/usr/bin/g++
/usr/bin/gcc
(exit_code=0) | terminal |
Role: You are a post-action environment simulator for a software-engineering editor and shell environment. The input action is already fixed; your role is to predict what the environment exposes next.
Input: Use the available goal, metadata, latest `###[Observation]`, action text, and history as state evidence. Do not... | **Turn 1 User**
###[Goal]
<uploaded_files>
/workspace/getmoto__moto__4.0
</uploaded_files>
I've uploaded a python code repository in the directory getmoto__moto__4.0. Consider the following PR description:
<pr_description>
Dynamodb update_item does not raise ValidationException when changing the partition key
Hi,
T... | ###[Observation]
EXECUTION RESULT of [str_replace_editor]:
Here's the files and directories up to 2 levels deep in /workspace/getmoto__moto__4.0/moto, excluding hidden items:
/workspace/getmoto__moto__4.0/moto
/workspace/getmoto__moto__4.0/moto/__init__.py
/workspace/getmoto__moto__4.0/moto/acm
/workspace/getmoto__moto... | swe |
Role: You are a post-action environment simulator for a terminal command-execution environment. The input action is already fixed; your role is to predict what the environment exposes next.
Input: Use the available goal, metadata, latest `###[Observation]`, action text, and history as state evidence. Do not add, skip,... | **Turn 1 User**
###[Goal]
Please solve this task:
Wake up! It's 3:00 AM and you've just been paged. The CI/CD pipeline for our core authentication service is failing during the configuration validation build step, completely blocking an emergency hotfix we need to deploy right now.
The build script `/home/user/app/va... | ###[Observation]
> > > > > > > Found 1 keys with gateway:
payment_gateway_843
(exit_code=0) | terminal |
Role: Serve as a language world model for a desktop accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state ... | **Turn 1 User**
###[Observation]
Computer ''
node=1 error '(-2147220991, \'An event was unable to invoke any of the subscribers\', (None, None, None, 0, None))'
###[Action]
First, use `mouse_move` to move the mouse pointer by a delta.
- This advances the task or changes state.
- Args: `delta`=[-482,379].
- Call:
```te... | ###[Observation]
Computer 'Word', b=-11,-11,1931,1019, raw_control='WindowControl'
node=1 toolbar '', raw_control='TitleBarControl', b=0,-8,1920,0
node=2 menubar 'System', raw_control='MenuBarControl', automation_id='MenuBar', b=0,0,33,33
node=3 menuitem 'System', raw_control='MenuItemControl', b=0,0,33,33
node... | os |
Role: Serve as a language world model for an MCP business-tool environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state snapshot. A... | **Turn 1 User**
###[Goal]
Cancel my currently active listing with ID 'LIST-77219' for the 'Obsidian Dragon Mount', return the item to my 'MythicLegends' game account inventory, and log the cancellation reason as 'Price too low'.
###[Action]
Use `list_tools` to list available MCP tools before selecting a specific tool.
... | ###[Observation]
{
"game_accounts": [
{
"id": 4,
"user_id": 1,
"game_id": 4,
"external_account_id": "MYL-ACC-420",
"display_name": "Mythic Main",
"linkage_status": "linked",
"created_at": "2026-05-19T13:01:55",
"updated_at": "2026-07-02T13:01:55"
}
]
} | mcp |
Role: Serve as a language world model for an MCP business-tool environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state snapshot. A... | **Turn 1 User**
###[Goal]
Schedule a distribution appointment for household ID 342 on 2025-01-10 at 10:30am in the 'North Side Pantry' location and generate a pickup order including 2 units of 'Pasta 1lb', 3 units of 'Canned Tomatoes 14oz', and 1 unit of 'Frozen Chicken 3lb'.
###[Action]
Use `list_tools` to list availa... | ###[Observation]
{
"items": [
{
"id": 4,
"name": "Pasta 1lb",
"category": "Grains",
"unit_size": "1 lb box",
"is_perishable": false,
"is_gluten_free": false,
"is_peanut_free": true,
"default_unit_weight_lbs": 1.0,
"created_at": "2024-01-01T08:15:00",
"up... | mcp |
Role: Serve as a language world model for a web search and page-browsing environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state s... | **Turn 1 User**
###[Goal]
Find the specific water depth in meters that marks the boundary between areas suitable for fixed foundation wind power and those requiring floating foundations according to the Global Wind Atlas (GWA) or the ESMAP-IFC Offshore Wind Development Program technical assessment criteria. I need the ... | ###[Observation]
Doc 2313
URL: https://www.esmap.org/esmap_offshorewind_techpotential_analysis_maps
L9: This methodology —which is described fully in the Going Global report— makes the following assumptions:
L10:
L11: * Regions with annual average 100-meter height wind speeds of greater than 7 meters/second are cons... | search |
Role: You are the simulator side of a language world model for a desktop accessibility-control environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the eviden... | **Turn 1 User**
###[Observation]
Computer 'Photos', b=433,169,1485,958, raw_control='WindowControl'
node=1 generic 'PopupHost', raw_control='PaneControl', b=834,772,1157,835
node=2 generic '', raw_control='PaneControl', b=444,169,1474,947
node=3 window 'Popup', raw_control='WindowControl', b=0,0,0,0
node=4 to... | ###[Observation]
Computer 'Photos', b=433,169,1485,958, raw_control='WindowControl'
node=1 generic 'Non Client Input Sink Window', raw_control='PaneControl', b=444,169,1474,241
node=2 generic 'AppWindow Custom Title Bar', raw_control='PaneControl', b=1258,169,1474,241
node=3 button 'Minimize', raw_control='ButtonCo... | os |
Role: You are the simulator side of a language world model for a desktop accessibility-control environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as the eviden... | **Turn 1 User**
###[Observation]
Computer 'Task ID 72 - Protected View - Word', b=-11,-11,1931,1019, raw_control='WindowControl'
node=1 generic 'DropShadowTop', raw_control='PaneControl', b=0,162,1920,171
node=2 generic 'MsoDockTop', raw_control='PaneControl', b=0,0,1920,162
node=3 toolbar '', raw_control='ToolBarC... | ###[Observation]
Computer 'Task ID 72 - Word', b=-11,-11,1931,1019, raw_control='WindowControl'
node=1 generic 'DropShadowTop', raw_control='PaneControl', b=0,267,1920,276
node=2 generic 'MsoDockTop', raw_control='PaneControl', b=0,0,1920,267
node=3 toolbar '', raw_control='ToolBarControl', b=0,0,1920,267
node=4... | os |
Role: Function as the environment model for a desktop accessibility-control environment. Given the action already chosen by another agent, infer the next visible observation without taking agency yourself.
Input: Read the goal or metadata, current observation, action, and history as a single interaction state. Only th... | **Turn 1 User**
###[Observation]
Computer 'Run', b=11,657,601,1004, raw_control='WindowControl'
node=1 image '', raw_control='ImageControl', automation_id='12297', b=40,736,88,784
node=2 text 'Type the name of a program, folder, document, or Internet resource, and Windows will open it for you.', raw_control='TextCont... | ###[Observation]
Computer 'Documents - File Explorer', b=370,64,1811,964, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=813,64,1800,115
node=2 generic 'Documents', raw_control='PaneControl', b=381,268,1800,953
node=3 generic '', raw_control='PaneControl', b=381,268,1800,953
node=4 g... | os |
Role: You are the simulator side of a language world model for a software-engineering editor and shell environment. The agent has already selected an action, and your job is to predict the environment response.
Input: Treat the goal or metadata, latest `###[Observation]`, current `###[Action]`, and earlier turns as th... | **Turn 1 User**
###[Goal]
<uploaded_files>
/workspace/iterative__dvc__0.92
</uploaded_files>
I've uploaded a python code repository in the directory iterative__dvc__0.92. Consider the following PR description:
<pr_description>
metrics diff with missing old value
If metrics file(s) has only no old version (it was just ... | ###[Observation]
EXECUTION RESULT of [str_replace_editor]:
Here's the files and directories up to 2 levels deep in /workspace/iterative__dvc__0.92/dvc, excluding hidden items:
/workspace/iterative__dvc__0.92/dvc
/workspace/iterative__dvc__0.92/dvc/__init__.py
/workspace/iterative__dvc__0.92/dvc/__main__.py
/workspace/i... | swe |
Role: You are a post-action environment simulator for a desktop accessibility-control environment. The input action is already fixed; your role is to predict what the environment exposes next.
Input: Use the available goal, metadata, latest `###[Observation]`, action text, and history as state evidence. Do not add, sk... | **Turn 1 User**
###[Observation]
Computer 'Task ID 43 - Excel', b=76,76,1516,893, raw_control='WindowControl'
node=1 generic '', raw_control='PaneControl', b=708,392,809,423
node=2 generic 'DropShadowTop', raw_control='PaneControl', b=87,345,1505,354
node=3 generic '', raw_control='PaneControl', b=87,849,1505,882
... | ###[Observation]
Computer '&Insert Statistic Chart', b=619,276,900,608, raw_control='MenuControl' | os |
Role: Serve as a language world model for a browser accessibility-control environment. Your output should be the environment state immediately after the recorded action, not the agent behavior before or after it.
Input: Combine the goal or task context, current observation, issued action, and prior turns into a state ... | **Turn 1 User**
###[Observation]
Web 'OpenStreetMap', focused
[44] banner '', visible
[45] heading 'OpenStreetMap logo OpenStreetMap', visible
[46] link 'OpenStreetMap logo OpenStreetMap', clickable, visible
[49] image 'OpenStreetMap logo', visible
StaticText ''
[51] navigation '', visible
[95] link ... | ###[Observation]
Web 'OpenStreetMap', focused
[44] banner '', visible
[45] heading 'OpenStreetMap logo OpenStreetMap', visible
[46] link 'OpenStreetMap logo OpenStreetMap', clickable, visible
[49] image 'OpenStreetMap logo', visible
StaticText ''
[51] navigation '', visible
[95] link 'Edit', clickabl... | web |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.