Dataset Viewer
Auto-converted to Parquet Duplicate
task_id
stringlengths
11
16
level
stringclasses
4 values
prompt
stringlengths
17
253
available_tools
listlengths
4
8
expected_trace
stringlengths
173
2.74k
expected_final_answer
stringlengths
18
1.74k
num_steps
int32
1
6
num_tools_offered
int32
4
8
category
stringclasses
9 values
pattern
stringlengths
0
29
L0_node_0001
L0_node
Can you add a specified number of days, hours, and/or minutes to a date. Returns the resulting date in ISO format? The parameters are date: 2025-02-14, minutes: 87.
[ "add_duration", "case_convert", "retrieve_memory", "data_filter" ]
{"steps": [{"step_id": "step_1", "tool_name": "add_duration", "arguments": {"date": "2025-02-14", "minutes": 87}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"result_date": "2025-02-14T01:27:00Z"}
1
4
external_services
L0_node_0002
L0_node
Can you add a specified number of days, hours, and/or minutes to a date. Returns the resulting date in ISO format? The parameters are date: 2026-03-01, hours: 72.
[ "add_duration", "json_extract", "format_number", "retrieve_memory" ]
{"steps": [{"step_id": "step_1", "tool_name": "add_duration", "arguments": {"date": "2026-03-01", "hours": 72}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"result_date": "2026-03-04T00:00:00Z"}
1
4
external_services
L0_node_0003
L0_node
Please encode a plain-text string to its Base64 representation.
[ "base64_encode", "knowledge_base_query", "business_days_between", "read_file" ]
{"steps": [{"step_id": "step_1", "tool_name": "base64_encode", "arguments": {"text": "The food industry is experiencing disruption from plant-based alternatives and lab-grown meat. Consumer preferences are shifting rapidly."}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"encoded": "VGhlIGZvb2QgaW5kdXN0cnkgaXMgZXhwZXJpZW5jaW5nIGRpc3J1cHRpb24gZnJvbSBwbGFudC1iYXNlZCBhbHRlcm5hdGl2ZXMgYW5kIGxhYi1ncm93biBtZWF0LiBDb25zdW1lciBwcmVmZXJlbmNlcyBhcmUgc2hpZnRpbmcgcmFwaWRseS4="}
1
4
text_processing
L0_node_0004
L0_node
Please encode a plain-text string to its Base64 representation.
[ "base64_encode", "standard_deviation", "percentile", "compare_texts" ]
{"steps": [{"step_id": "step_1", "tool_name": "base64_encode", "arguments": {"text": "Quantum computing promises to revolutionize how we process information. Unlike classical bits, quantum bits can exist in multiple states simultaneously."}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"encoded": "UXVhbnR1bSBjb21wdXRpbmcgcHJvbWlzZXMgdG8gcmV2b2x1dGlvbml6ZSBob3cgd2UgcHJvY2VzcyBpbmZvcm1hdGlvbi4gVW5saWtlIGNsYXNzaWNhbCBiaXRzLCBxdWFudHVtIGJpdHMgY2FuIGV4aXN0IGluIG11bHRpcGxlIHN0YXRlcyBzaW11bHRhbmVvdXNseS4="}
1
4
text_processing
L0_node_0005
L0_node
Please count the number of business days (Monday–Friday) between two dates, exclusive of the start date and inclusive of the end date. Also returns the total calendar days — specifically, start date: 2026-02-01, end date: 2026-02-14.
[ "business_days_between", "list_memories", "round_number", "regex_match" ]
{"steps": [{"step_id": "step_1", "tool_name": "business_days_between", "arguments": {"start_date": "2026-02-01", "end_date": "2026-02-14"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"business_days": 10, "total_days": 13}
1
4
external_services
L0_node_0006
L0_node
Please calculate the difference between two dates in days, weeks, months, or years — specifically, date1: 2025-11-15, date2: 2026-02-01.
[ "calculate_date_diff", "string_replace", "clamp_value", "get_weather" ]
{"steps": [{"step_id": "step_1", "tool_name": "calculate_date_diff", "arguments": {"date1": "2025-11-15", "date2": "2026-02-01"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"date1": "2025-11-15", "date2": "2026-02-01", "days": 78, "weeks": 11.1, "months": 2.6, "years": 0.21, "date1_is_before": true}
1
4
time_scheduling
L0_node_0007
L0_node
I need you to calculate the difference between two dates in days, weeks, months, or years. Here are the details: date1: 2026-04-01, date2: 2026-02-14.
[ "calculate_date_diff", "compare_texts", "string_replace", "get_exchange_rate" ]
{"steps": [{"step_id": "step_1", "tool_name": "calculate_date_diff", "arguments": {"date1": "2026-04-01", "date2": "2026-02-14"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"date1": "2026-04-01", "date2": "2026-02-14", "days": 46, "weeks": 6.6, "months": 1.5, "years": 0.13, "date1_is_before": false}
1
4
time_scheduling
L0_node_0008
L0_node
Compute the value of 177 - 61.
[ "calculator", "string_replace", "validate_email", "get_session_context" ]
{"steps": [{"step_id": "step_1", "tool_name": "calculator", "arguments": {"expression": "177 - 61"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"expression": "177 - 61", "result": 116.0}
1
4
computation
L0_node_0009
L0_node
What is 234 - 89?
[ "calculator", "transform_format", "schedule_meeting", "extract_domain" ]
{"steps": [{"step_id": "step_1", "tool_name": "calculator", "arguments": {"expression": "234 - 89"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"expression": "234 - 89", "result": 145.0}
1
4
computation
L0_node_0010
L0_node
I need you to convert text to a different case: upper, lower, title, camelCase, or snake_case. Here are the details: target case: lower.
[ "case_convert", "get_session_context", "slugify", "moving_average" ]
{"steps": [{"step_id": "step_1", "tool_name": "case_convert", "arguments": {"text": "The global economy is undergoing a period of significant transformation. Digital technologies are reshaping traditional business models.", "target_case": "lower"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"converted": "the global economy is undergoing a period of significant transformation. digital technologies are reshaping traditional business models."}
1
4
text_processing
L0_node_0011
L0_node
I need you to convert text to a different case: upper, lower, title, camelCase, or snake_case. Here are the details: target case: camel.
[ "case_convert", "next_occurrence", "database_query", "time_since" ]
{"steps": [{"step_id": "step_1", "tool_name": "case_convert", "arguments": {"text": "Fusion energy research has achieved significant milestones. Commercial fusion power could provide virtually unlimited clean energy.", "target_case": "camel"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"converted": "fusionEnergyResearchHasAchievedSignificantMilestones.CommercialFusionPowerCouldProvideVirtuallyUnlimitedCleanEnergy."}
1
4
text_processing
L0_node_0012
L0_node
Can you check whether a URL is reachable and return its HTTP status code and estimated response time? The parameters are url: https://news.com.
[ "check_url_status", "case_convert", "extract_entities", "text_to_number" ]
{"steps": [{"step_id": "step_1", "tool_name": "check_url_status", "arguments": {"url": "https://news.com"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"url": "https://news.com", "status": "reachable", "status_code": 200, "response_time_ms": 485}
1
4
information_retrieval
L0_node_0013
L0_node
Check whether a URL is reachable and return its HTTP status code and estimated response time with the following: url: https://blog.dev.
[ "check_url_status", "generate_report", "data_filter", "http_request" ]
{"steps": [{"step_id": "step_1", "tool_name": "check_url_status", "arguments": {"url": "https://blog.dev"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"url": "https://blog.dev", "status": "client_error", "status_code": 403, "response_time_ms": 200}
1
4
information_retrieval
L0_node_0014
L0_node
Clamp a number so it falls within the specified minimum and maximum range. Returns the clamped value and whether clamping was applied with the following: value: 86.25, min val: 16.13, max val: 16.84.
[ "clamp_value", "word_count", "number_to_text", "compare_texts" ]
{"steps": [{"step_id": "step_1", "tool_name": "clamp_value", "arguments": {"value": 86.25, "min_val": 16.13, "max_val": 16.84}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"clamped": 16.84, "was_clamped": true}
1
4
computation
L0_node_0015
L0_node
Please classify a piece of text into one or more predefined categories based on its content.
[ "classify_text", "word_count", "compress_data", "parse_html" ]
{"steps": [{"step_id": "step_1", "tool_name": "classify_text", "arguments": {"text": "Microplastics have been found in every environment on Earth. Research into their health effects is intensifying.", "categories": ["item_0", "item_1"]}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"text_preview": "microplastics have been found in every environment on earth. research into their health effects is i...", "predicted_category": "item_0", "confidence": 0.0, "all_scores": {"item_0": 0.0, "item_1": 0.0}}
1
4
text_processing
L0_node_0016
L0_node
Classify a piece of text into one or more predefined categories based on its content.
[ "classify_text", "number_to_text", "get_weather", "extract_numbers" ]
{"steps": [{"step_id": "step_1", "tool_name": "classify_text", "arguments": {"text": "Quantum computing promises to revolutionize how we process information. Unlike classical bits, quantum bits can exist in multiple states simultaneously.", "categories": ["item_0", "item_1", "item_2", "item_3"]}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"text_preview": "quantum computing promises to revolutionize how we process information. unlike classical bits, quant...", "predicted_category": "item_0", "confidence": 0.0, "all_scores": {"item_0": 0.0, "item_1": 0.0, "item_2": 0.0, "item_3": 0.0}}
1
4
text_processing
L0_node_0017
L0_node
I need you to compare two pieces of text and return their similarity score, common keywords, and differences.
[ "compare_texts", "add_duration", "regex_match", "round_number" ]
{"steps": [{"step_id": "step_1", "tool_name": "compare_texts", "arguments": {"text1": "Renewable energy adoption is accelerating globally. Solar and wind power costs have decreased dramatically over the past decade.", "text2": "Biodegradable materials are replacing traditional plastics in packaging. Consumer demand is driving innovation in sustainable materials."}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"similarity_score": 0.032, "common_words_count": 1, "text1_unique_words": 17, "text2_unique_words": 13, "common_keywords": ["is"], "text1_length": 128, "text2_length": 136}
1
4
text_processing
L0_node_0018
L0_node
Compare two pieces of text and return their similarity score, common keywords, and differences.
[ "compare_texts", "normalize_data", "create_invoice", "create_notification" ]
{"steps": [{"step_id": "step_1", "tool_name": "compare_texts", "arguments": {"text1": "Biodiversity loss is accelerating at an unprecedented rate. Conservation efforts must be scaled up significantly to protect endangered species.", "text2": "The metaverse concept is evolving from gaming to business applications. Virtual reality meetings and digital workspaces are becoming mainstream."}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"similarity_score": 0.056, "common_words_count": 2, "text1_unique_words": 17, "text2_unique_words": 17, "common_keywords": ["is", "to"], "text1_length": 143, "text2_length": 144}
1
4
text_processing
L0_node_0019
L0_node
Compress a data string using a specified compression algorithm. Returns the compressed size, original size, and compression ratio with the following: data: sample_data_246, algorithm: gzip.
[ "compress_data", "transform_format", "summarize_text", "data_aggregate" ]
{"steps": [{"step_id": "step_1", "tool_name": "compress_data", "arguments": {"data": "sample_data_246", "algorithm": "gzip"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"compressed_size": 23, "original_size": 15, "ratio": 1.5333, "algorithm": "gzip"}
1
4
communication
L0_node_0020
L0_node
Can you compress a data string using a specified compression algorithm. Returns the compressed size, original size, and compression ratio? The parameters are data: sample_data_779, algorithm: lz4.
[ "compress_data", "check_url_status", "parse_date", "base64_encode" ]
{"steps": [{"step_id": "step_1", "tool_name": "compress_data", "arguments": {"data": "sample_data_779", "algorithm": "lz4"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"compressed_size": 23, "original_size": 15, "ratio": 1.5333, "algorithm": "lz4"}
1
4
communication
L0_node_0021
L0_node
Create a calendar event with a title, date, duration, and optional list of attendees. Returns a confirmation with the event ID with the following: title: Retrospective, date: 2026-02-14, duration minutes: 91.
[ "create_calendar_event", "data_sort", "schedule_meeting", "correlation" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_calendar_event", "arguments": {"title": "Retrospective", "date": "2026-02-14", "duration_minutes": 91, "attendees": ["item_0", "item_1", "item_2"]}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"event_id": "evt_2c4e30a2bd53", "title": "Retrospective", "date": "2026-02-14", "duration_minutes": 91, "attendees": ["item_0", "item_1", "item_2"], "confirmation": "Calendar event 'Retrospective' created for 2026-02-14."}
1
4
communication
L0_node_0022
L0_node
Can you create a calendar event with a title, date, duration, and optional list of attendees. Returns a confirmation with the event ID? The parameters are title: Marketing Sync, date: 2026-01-01, duration minutes: 93.
[ "create_calendar_event", "merge_data", "ip_geolocation", "create_contact" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_calendar_event", "arguments": {"title": "Marketing Sync", "date": "2026-01-01", "duration_minutes": 93, "attendees": ["item_0", "item_1", "item_2", "item_3", "item_4"]}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"event_id": "evt_7f9fe9e7a5d3", "title": "Marketing Sync", "date": "2026-01-01", "duration_minutes": 93, "attendees": ["item_0", "item_1", "item_2", "item_3", "item_4"], "confirmation": "Calendar event 'Marketing Sync' created for 2026-01-01."}
1
4
communication
L0_node_0023
L0_node
Can you create a new contact entry with a name, email address, and optional phone number? The parameters are name: sample_name_675, email: nick@marketing.brand, phone: sample_phone_670.
[ "create_contact", "url_parse", "format_date", "statistical_analysis" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_contact", "arguments": {"name": "sample_name_675", "email": "nick@marketing.brand", "phone": "sample_phone_670"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"contact_id": "con_6e199e2e4173", "name": "sample_name_675", "email": "nick@marketing.brand", "phone": "sample_phone_670", "confirmation": "Contact 'sample_name_675' created."}
1
4
communication
L0_node_0024
L0_node
I need you to create a new contact entry with a name, email address, and optional phone number. Here are the details: name: sample_name_255, email: olivia@sales.deal.
[ "create_contact", "get_location_info", "knowledge_base_query", "create_calendar_event" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_contact", "arguments": {"name": "sample_name_255", "email": "olivia@sales.deal"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"contact_id": "con_259c4c78c73c", "name": "sample_name_255", "email": "olivia@sales.deal", "phone": null, "confirmation": "Contact 'sample_name_255' created."}
1
4
communication
L0_node_0025
L0_node
Can you create a notification or alert with a title and message, optionally with a priority level? The parameters are title: Cross-Team Alignment, message: CI/CD pipeline completed in 3m 42s.
[ "create_notification", "spell_check", "web_page_fetch", "ip_geolocation" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_notification", "arguments": {"title": "Cross-Team Alignment", "message": "CI/CD pipeline completed in 3m 42s"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"status": "created", "notification_id": "notif_25a181a23638", "title": "Cross-Team Alignment", "priority": "normal", "timestamp": "2026-02-22T12:00:00"}
1
4
communication
L0_node_0026
L0_node
Create a notification or alert with a title and message, optionally with a priority level with the following: title: Onboarding Orientation, message: Automated backup verification passed, priority: normal.
[ "create_notification", "round_number", "write_file", "data_sort" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_notification", "arguments": {"title": "Onboarding Orientation", "message": "Automated backup verification passed", "priority": "normal"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"status": "created", "notification_id": "notif_13f776b4ab23", "title": "Onboarding Orientation", "priority": "normal", "timestamp": "2026-02-22T12:00:00"}
1
4
communication
L0_node_0027
L0_node
I need you to create a spreadsheet with a title, column headers, and data rows. Returns a spreadsheet ID and row count. Here are the details: title: Budget Review.
[ "create_spreadsheet", "statistical_analysis", "log_event", "data_aggregate" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_spreadsheet", "arguments": {"title": "Budget Review", "headers": ["item_0", "item_1", "item_2", "item_3", "item_4"], "rows": ["item_0", "item_1", "item_2"]}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"spreadsheet_id": "sht_757ab09676fa", "title": "Budget Review", "columns": 5, "row_count": 3, "headers": ["item_0", "item_1", "item_2", "item_3", "item_4"], "confirmation": "Spreadsheet 'Budget Review' created with 3 rows and 5 columns."}
1
4
communication
L0_node_0028
L0_node
Create a spreadsheet with a title, column headers, and data rows. Returns a spreadsheet ID and row count with the following: title: Marketing Sync.
[ "create_spreadsheet", "get_weekday", "summarize_text", "merge_data" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_spreadsheet", "arguments": {"title": "Marketing Sync", "headers": ["item_0", "item_1", "item_2"], "rows": ["item_0", "item_1", "item_2", "item_3", "item_4"]}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"spreadsheet_id": "sht_c38040eb64a3", "title": "Marketing Sync", "columns": 3, "row_count": 5, "headers": ["item_0", "item_1", "item_2"], "confirmation": "Spreadsheet 'Marketing Sync' created with 5 rows and 3 columns."}
1
4
communication
L0_node_0029
L0_node
Please create a task or to-do item with a title, description, optional due date, and priority — specifically, title: Marketing Sync, description: sample_description_210, due date: 2026-05-01.
[ "create_task", "url_parse", "min_max", "create_calendar_event" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_task", "arguments": {"title": "Marketing Sync", "description": "sample_description_210", "due_date": "2026-05-01"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"status": "created", "task_id": "task_dcbecb0e6809", "title": "Marketing Sync", "description": "sample_description_210", "due_date": "2026-05-01", "priority": "medium", "created_at": "2026-02-22T12:00:00"}
1
4
communication
L0_node_0030
L0_node
Please create a task or to-do item with a title, description, optional due date, and priority — specifically, title: Vendor Evaluation, description: sample_description_551.
[ "create_task", "standard_deviation", "execute_python", "web_search" ]
{"steps": [{"step_id": "step_1", "tool_name": "create_task", "arguments": {"title": "Vendor Evaluation", "description": "sample_description_551"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"status": "created", "task_id": "task_69df17868447", "title": "Vendor Evaluation", "description": "sample_description_551", "due_date": null, "priority": "medium", "created_at": "2026-02-22T12:00:00"}
1
4
communication
L0_node_0031
L0_node
Aggregate a list of records by a group key, computing sum, average, count, min, or max for a specified value field with the following: group by: sample_group_by_721, value field: sample_value_field_542.
[ "data_aggregate", "generate_url", "list_memories", "min_max" ]
{"steps": [{"step_id": "step_1", "tool_name": "data_aggregate", "arguments": {"items": ["item_0", "item_1", "item_2", "item_3"], "group_by": "sample_group_by_721", "value_field": "sample_value_field_542"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"operation": "sum", "group_by": "sample_group_by_721", "value_field": "sample_value_field_542", "groups": {}, "total_records": 4, "total_groups": 0}
1
4
computation
L0_node_0032
L0_node
Please aggregate a list of records by a group key, computing sum, average, count, min, or max for a specified value field — specifically, group by: sample_group_by_761, value field: sample_value_field_289, operation: count.
[ "data_aggregate", "read_file", "word_count", "string_replace" ]
{"steps": [{"step_id": "step_1", "tool_name": "data_aggregate", "arguments": {"items": ["item_0", "item_1", "item_2"], "group_by": "sample_group_by_761", "value_field": "sample_value_field_289", "operation": "count"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"operation": "count", "group_by": "sample_group_by_761", "value_field": "sample_value_field_289", "groups": {}, "total_records": 3, "total_groups": 0}
1
4
computation
L0_node_0033
L0_node
Filter a list of items based on a condition. For numbers: greater_than, less_than, equals. For strings: contains, starts_with, ends_with with the following: condition: equals, value: sample_value_650.
[ "data_filter", "number_to_text", "lookup_entity", "calculator" ]
{"steps": [{"step_id": "step_1", "tool_name": "data_filter", "arguments": {"items": ["item_0", "item_1", "item_2", "item_3", "item_4"], "condition": "equals", "value": "sample_value_650"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"filtered": [], "count": 0, "original_count": 5, "condition": "equals", "value": "sample_value_650"}
1
4
computation
L0_node_0034
L0_node
Can you filter a list of items based on a condition. For numbers: greater_than, less_than, equals. For strings: contains, starts_with, ends_with? The parameters are condition: ends_with, value: sample_value_166.
[ "data_filter", "case_convert", "mask_pii", "list_memories" ]
{"steps": [{"step_id": "step_1", "tool_name": "data_filter", "arguments": {"items": ["item_0", "item_1", "item_2", "item_3"], "condition": "ends_with", "value": "sample_value_166"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"filtered": [], "count": 0, "original_count": 4, "condition": "ends_with", "value": "sample_value_166"}
1
4
computation
L0_node_0035
L0_node
I need you to sort a list of items (numbers or strings) in ascending or descending order. Here are the details: key: project_notes.
[ "data_sort", "transcribe_audio", "sentiment_analysis", "list_memories" ]
{"steps": [{"step_id": "step_1", "tool_name": "data_sort", "arguments": {"items": ["item_0", "item_1"], "key": "project_notes"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"sorted": ["item_0", "item_1"], "order": "ascending", "count": 2}
1
4
computation
L0_node_0036
L0_node
Sort a list of items (numbers or strings) in ascending or descending order with the following: key: recipe_collection.
[ "data_sort", "word_count", "web_search", "base64_decode" ]
{"steps": [{"step_id": "step_1", "tool_name": "data_sort", "arguments": {"items": ["item_0", "item_1", "item_2", "item_3", "item_4"], "key": "recipe_collection"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"sorted": ["item_0", "item_1", "item_2", "item_3", "item_4"], "order": "ascending", "count": 5}
1
4
computation
L0_node_0037
L0_node
Please query a structured database table. Supports filtering, sorting, and aggregation on predefined datasets (countries, cities, movies, books) — specifically, table: movies, filter field: sample_filter_field_768, filter value: sample_filter_value_903.
[ "database_query", "case_convert", "extract_domain", "merge_data" ]
{"steps": [{"step_id": "step_1", "tool_name": "database_query", "arguments": {"table": "movies", "filter_field": "sample_filter_field_768", "filter_value": "sample_filter_value_903", "sort_by": "sample_sort_by_802", "limit": 86}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"table": "movies", "results": [], "count": 0, "error": "Table not found"}
1
4
information_retrieval
L0_node_0038
L0_node
Please query a structured database table. Supports filtering, sorting, and aggregation on predefined datasets (countries, cities, movies, books) — specifically, table: cities, filter field: sample_filter_field_880, filter op: greater_than.
[ "database_query", "linear_regression", "lookup_entity", "data_sort" ]
{"steps": [{"step_id": "step_1", "tool_name": "database_query", "arguments": {"table": "cities", "filter_field": "sample_filter_field_880", "filter_op": "greater_than", "filter_value": "sample_filter_value_841", "sort_by": "sample_sort_by_676"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"table": "cities", "results": [], "count": 0}
1
4
information_retrieval
L0_node_0039
L0_node
Detect the language of the input text. Uses common function-word heuristics to identify English, French, German, Spanish, Italian, Portuguese, and Dutch.
[ "detect_language", "hash_text", "store_memory", "correlation" ]
{"steps": [{"step_id": "step_1", "tool_name": "detect_language", "arguments": {"text": "The education sector is embracing technology-enhanced learning. Online platforms and AI tutors are making education more accessible."}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"language": "English", "language_code": "en", "confidence": 0.8}
1
4
text_processing
L0_node_0040
L0_node
Can you detect the language of the input text. Uses common function-word heuristics to identify English, French, German, Spanish, Italian, Portuguese, and Dutch?
[ "detect_language", "database_query", "schedule_meeting", "add_duration" ]
{"steps": [{"step_id": "step_1", "tool_name": "detect_language", "arguments": {"text": "Gene editing technologies like CRISPR offer enormous potential for treating genetic diseases. Ethical considerations remain important."}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"language": "English", "language_code": "en", "confidence": 1.0}
1
4
text_processing
L0_node_0041
L0_node
Can you perform a DNS lookup for a domain and return the resolved IP address and record type? The parameters are domain: sample_domain_447.
[ "dns_lookup", "compare_texts", "log_event", "is_business_day" ]
{"steps": [{"step_id": "step_1", "tool_name": "dns_lookup", "arguments": {"domain": "sample_domain_447"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"domain": "sample_domain_447", "ip_address": "117.137.50.116", "record_type": "A", "ttl_seconds": 3424}
1
4
information_retrieval
L0_node_0042
L0_node
Perform a DNS lookup for a domain and return the resolved IP address and record type with the following: domain: sample_domain_261.
[ "dns_lookup", "calculator", "format_number", "number_to_text" ]
{"steps": [{"step_id": "step_1", "tool_name": "dns_lookup", "arguments": {"domain": "sample_domain_261"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"domain": "sample_domain_261", "ip_address": "77.12.18.40", "record_type": "A", "ttl_seconds": 2784}
1
4
information_retrieval
L0_node_0043
L0_node
Please uRL encode or decode a text string. Encodes special characters for safe use in URLs, or decodes percent-encoded strings back to readable text — specifically, action: encode.
[ "encode_url", "clamp_value", "number_to_text", "data_sort" ]
{"steps": [{"step_id": "step_1", "tool_name": "encode_url", "arguments": {"text": "Neuroscience breakthroughs are improving our understanding of brain function. Brain-computer interfaces could transform healthcare.", "action": "encode"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"result": "Neuroscience%20breakthroughs%20are%20improving%20our%20understanding%20of%20brain%20function.%20Brain-computer%20interfaces%20could%20transform%20healthcare."}
1
4
information_retrieval
L0_node_0044
L0_node
Please uRL encode or decode a text string. Encodes special characters for safe use in URLs, or decodes percent-encoded strings back to readable text — specifically, action: encode.
[ "encode_url", "truncate_text", "round_number", "send_webhook" ]
{"steps": [{"step_id": "step_1", "tool_name": "encode_url", "arguments": {"text": "The Internet of Things connects billions of devices worldwide. Smart home technology is becoming standard in new construction.", "action": "encode"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"result": "The%20Internet%20of%20Things%20connects%20billions%20of%20devices%20worldwide.%20Smart%20home%20technology%20is%20becoming%20standard%20in%20new%20construction."}
1
4
information_retrieval
L0_node_0045
L0_node
Can you encrypt a text string using a specified encryption method. Returns the simulated encrypted output as a base64 string? The parameters are method: rsa.
[ "encrypt_text", "list_memories", "parse_date", "translate_text" ]
{"steps": [{"step_id": "step_1", "tool_name": "encrypt_text", "arguments": {"text": "The education sector is embracing technology-enhanced learning. Online platforms and AI tutors are making education more accessible.", "method": "rsa"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"encrypted": "sBIJK3hty3oxA54csWwdY9xTiUBfjTxcc1ypiAaYwfk=", "method": "rsa", "original_length": 132, "encrypted_length": 44}
1
4
communication
L0_node_0046
L0_node
Can you encrypt a text string using a specified encryption method. Returns the simulated encrypted output as a base64 string? The parameters are method: aes256.
[ "encrypt_text", "convert_timezone", "summarize_text", "round_number" ]
{"steps": [{"step_id": "step_1", "tool_name": "encrypt_text", "arguments": {"text": "Neuroscience breakthroughs are improving our understanding of brain function. Brain-computer interfaces could transform healthcare.", "method": "aes256"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"encrypted": "nrRSS4U2c37qWog1HIA1B3UV/LwKKukaaUz4xAtdLzk=", "method": "aes256", "original_length": 131, "encrypted_length": 44}
1
4
communication
L0_node_0047
L0_node
Can you extract the domain, subdomain, and top-level domain (TLD) from a given URL? The parameters are url: https://news.com.
[ "extract_domain", "generate_summary_stats", "base64_decode", "read_file" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_domain", "arguments": {"url": "https://news.com"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"domain": "news", "subdomain": "", "tld": "com"}
1
4
information_retrieval
L0_node_0048
L0_node
Please extract the domain, subdomain, and top-level domain (TLD) from a given URL — specifically, url: https://blog.dev.
[ "extract_domain", "data_aggregate", "lookup_entity", "log_event" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_domain", "arguments": {"url": "https://blog.dev"}, "depends_on": [], "output_key": null}], "final_answer_source": "step_1"}
{"domain": "blog", "subdomain": "", "tld": "dev"}
1
4
information_retrieval
L1_chain_0049
L1_chain
Check the weather in Berlin and convert the temperature to Fahrenheit.
[ "get_weather", "unit_convert", "word_count", "normalize_data", "transform_format" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Berlin"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "unit_convert", "arguments": {"value": 36, "from_unit": "celsius", "to_unit": "fahrenheit"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_value": 36.0, "from_unit": "celsius", "converted_value": 96.8, "to_unit": "fahrenheit"}
2
5
chain
retrieve-transform
L1_chain_0050
L1_chain
Check the weather in Cairo and convert the temperature to Fahrenheit.
[ "get_weather", "unit_convert", "create_task", "calculator", "database_query" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Cairo"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "unit_convert", "arguments": {"value": 33, "from_unit": "celsius", "to_unit": "fahrenheit"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_value": 33.0, "from_unit": "celsius", "converted_value": 91.4, "to_unit": "fahrenheit"}
2
5
chain
retrieve-transform
L1_chain_0051
L1_chain
What is the temperature in Helsinki in Fahrenheit?
[ "get_weather", "unit_convert", "set_reminder", "number_to_text", "string_replace" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Helsinki"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "unit_convert", "arguments": {"value": 30, "from_unit": "celsius", "to_unit": "fahrenheit"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_value": 30.0, "from_unit": "celsius", "converted_value": 86.0, "to_unit": "fahrenheit"}
2
5
chain
retrieve-transform
L1_chain_0052
L1_chain
Find information about "digital twin technology" and give me a summary.
[ "web_search", "summarize_text", "split_text", "slugify", "spell_check" ]
{"steps": [{"step_id": "step_1", "tool_name": "web_search", "arguments": {"query": "digital twin technology"}, "depends_on": [], "output_key": "results"}, {"step_id": "step_2", "tool_name": "summarize_text", "arguments": {"text": "[{\"title\": \"Digital Twin Technology \\u2014 Wikipedia Overview\", \"url\": \"https://en.wikipedia.org/article/40290\", \"snippet\": \"A comprehensive overview of digital twin technology covering key concepts, recent developments, and practical applications.\", \"source\": \"Wikipedia\"}, {\"title\": \"Digital Twin Technology \\u2014 Academic Analysis\", \"url\": \"https://sciencedirect.com/article/35383\", \"snippet\": \"Expert guide to digital twin technology with detailed explanations, examples, and best practices for "}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_length": 500, "summary": "[{\"title\": \"Digital Twin Technology \\u2014 Wikipedia Overview\", \"url\": \"https://en.wikipedia.org/article/40290\", \"snippet\": \"A comprehensive overview of digital twin technology covering key concepts, recent developments, and practical applications.\", \"source\": \"Wikipedia\"}, {\"title\": \"Digital Twin Technology \\u2014 Academic Analysis\", \"url\": \"https://sciencedirect.com/article/35383\", \"snippet\": \"Expert guide to digital twin technology with detailed explanations, examples, and best practices for", "summary_length": 499, "sentences_used": 1, "compression_ratio": 1.0}
2
5
chain
retrieve-process
L1_chain_0053
L1_chain
Search the web for "autonomous vehicle regulations", then summarize the findings.
[ "web_search", "summarize_text", "merge_data", "next_occurrence", "base64_encode" ]
{"steps": [{"step_id": "step_1", "tool_name": "web_search", "arguments": {"query": "autonomous vehicle regulations"}, "depends_on": [], "output_key": "results"}, {"step_id": "step_2", "tool_name": "summarize_text", "arguments": {"text": "[{\"title\": \"Autonomous Vehicle Regulations \\u2014 News Overview\", \"url\": \"https://stackoverflow.com/article/41076\", \"snippet\": \"Latest research findings on autonomous vehicle regulations from leading institutions and peer-reviewed publications.\", \"source\": \"News\"}, {\"title\": \"Autonomous Vehicle Regulations \\u2014 News Analysis\", \"url\": \"https://stackoverflow.com/article/93556\", \"snippet\": \"Latest research findings on autonomous vehicle regulations from leading institutions and peer-reviewed publ"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_length": 500, "summary": "[{\"title\": \"Autonomous Vehicle Regulations \\u2014 News Overview\", \"url\": \"https://stackoverflow.com/article/41076\", \"snippet\": \"Latest research findings on autonomous vehicle regulations from leading institutions and peer-reviewed publications.\", \"source\": \"News\"}, {\"title\": \"Autonomous Vehicle Regulations \\u2014 News Analysis\", \"url\": \"https://stackoverflow.com/article/93556\", \"snippet\": \"Latest research findings on autonomous vehicle regulations from leading institutions and peer-reviewed publ", "summary_length": 500, "sentences_used": 1, "compression_ratio": 1.0}
2
5
chain
retrieve-process
L1_chain_0054
L1_chain
Search for "water purification methods" and summarize the results.
[ "web_search", "summarize_text", "extract_links", "dns_lookup", "add_duration" ]
{"steps": [{"step_id": "step_1", "tool_name": "web_search", "arguments": {"query": "water purification methods"}, "depends_on": [], "output_key": "results"}, {"step_id": "step_2", "tool_name": "summarize_text", "arguments": {"text": "[{\"title\": \"Water Purification Methods \\u2014 Blog Overview\", \"url\": \"https://nytimes.com/article/12187\", \"snippet\": \"water purification methods: analysis of current trends, methodologies, and future directions in the field.\", \"source\": \"Blog\"}, {\"title\": \"Water Purification Methods \\u2014 Wikipedia Analysis\", \"url\": \"https://github.com/article/59325\", \"snippet\": \"A comprehensive overview of water purification methods covering key concepts, recent developments, and practical applications.\", \"sou"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_length": 500, "summary": "[{\"title\": \"Water Purification Methods \\u2014 Blog Overview\", \"url\": \"https://nytimes.com/article/12187\", \"snippet\": \"water purification methods: analysis of current trends, methodologies, and future directions in the field.\", \"source\": \"Blog\"}, {\"title\": \"Water Purification Methods \\u2014 Wikipedia Analysis\", \"url\": \"https://github.com/article/59325\", \"snippet\": \"A comprehensive overview of water purification methods covering key concepts, recent developments, and practical applications.\", \"sou", "summary_length": 500, "sentences_used": 1, "compression_ratio": 1.0}
2
5
chain
retrieve-process
L1_chain_0055
L1_chain
What is V's stock price in NZD?
[ "get_stock_price", "get_exchange_rate", "translate_text", "extract_domain", "slugify" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_stock_price", "arguments": {"symbol": "V"}, "depends_on": [], "output_key": "stock"}, {"step_id": "step_2", "tool_name": "get_exchange_rate", "arguments": {"from_currency": "USD", "to_currency": "NZD", "amount": 113}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"from_currency": "USD", "to_currency": "NZD", "rate": 1.0, "amount": 113.0, "converted_amount": 113.0}
2
5
chain
retrieve-convert
L1_chain_0056
L1_chain
What is UBER's stock price in SEK?
[ "get_stock_price", "get_exchange_rate", "summarize_text", "database_query", "get_weekday" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_stock_price", "arguments": {"symbol": "UBER"}, "depends_on": [], "output_key": "stock"}, {"step_id": "step_2", "tool_name": "get_exchange_rate", "arguments": {"from_currency": "USD", "to_currency": "SEK", "amount": 133}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"from_currency": "USD", "to_currency": "SEK", "rate": 10.42, "amount": 133.0, "converted_amount": 1385.86}
2
5
chain
retrieve-convert
L1_chain_0057
L1_chain
Get the price of SQ and convert it to CAD.
[ "get_stock_price", "get_exchange_rate", "hash_text", "store_memory", "summarize_text" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_stock_price", "arguments": {"symbol": "SQ"}, "depends_on": [], "output_key": "stock"}, {"step_id": "step_2", "tool_name": "get_exchange_rate", "arguments": {"from_currency": "USD", "to_currency": "CAD", "amount": 539}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"from_currency": "USD", "to_currency": "CAD", "rate": 1.36, "amount": 539.0, "converted_amount": 733.04}
2
5
chain
retrieve-convert
L1_chain_0058
L1_chain
Find named entities in "Rosalind Franklin's X-ray crystallography work was done at King's College London..." and determine the sentiment.
[ "extract_entities", "sentiment_analysis", "extract_links", "generate_summary_stats", "calculate_date_diff" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_entities", "arguments": {"text": "Rosalind Franklin's X-ray crystallography work was done at King's College London in 1952."}, "depends_on": [], "output_key": "entities"}, {"step_id": "step_2", "tool_name": "sentiment_analysis", "arguments": {"text": "Rosalind Franklin's X-ray crystallography work was done at King's College London in 1952."}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"text_length": 89, "sentiment": "neutral", "confidence": 0.5, "positive_signals": 0, "negative_signals": 0}
2
5
chain
analyze-classify
L1_chain_0059
L1_chain
Identify entities in "Final exam schedule: email registrar@university.edu or visit room 301 in Admin B...", then classify its sentiment.
[ "extract_entities", "sentiment_analysis", "send_webhook", "text_to_number", "retrieve_memory" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_entities", "arguments": {"text": "Final exam schedule: email registrar@university.edu or visit room 301 in Admin Building."}, "depends_on": [], "output_key": "entities"}, {"step_id": "step_2", "tool_name": "sentiment_analysis", "arguments": {"text": "Final exam schedule: email registrar@university.edu or visit room 301 in Admin Building."}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"text_length": 88, "sentiment": "neutral", "confidence": 0.5, "positive_signals": 0, "negative_signals": 0}
2
5
chain
analyze-classify
L1_chain_0060
L1_chain
Extract entities from "Registration closes on 06/30/2026. Contact events@conference.org for more inform..." and analyze the sentiment.
[ "extract_entities", "sentiment_analysis", "parse_html", "set_reminder", "mask_pii" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_entities", "arguments": {"text": "Registration closes on 06/30/2026. Contact events@conference.org for more information."}, "depends_on": [], "output_key": "entities"}, {"step_id": "step_2", "tool_name": "sentiment_analysis", "arguments": {"text": "Registration closes on 06/30/2026. Contact events@conference.org for more information."}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"text_length": 86, "sentiment": "neutral", "confidence": 0.5, "positive_signals": 0, "negative_signals": 0}
2
5
chain
analyze-classify
L1_chain_0061
L1_chain
Look up weather in Manila, then email the results to yuki@international.global.
[ "get_weather", "send_email", "send_webhook", "string_replace", "web_search" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Manila"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "send_email", "arguments": {"to": "yuki@international.global", "subject": "Weather in Manila", "body": "{\"city\": \"Manila\", \"temperature_celsius\": 26, \"humidity_percent\": 31, \"wind_speed_kmh\": 41, \"condition\": \"foggy\", \"feels_like_celsius\": 24}"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"status": "sent", "message_id": "msg_a7ab2625ae69", "to": "yuki@international.global", "subject": "Weather in Manila", "body_preview": "{\"city\": \"Manila\", \"temperature_celsius\": 26, \"humidity_percent\": 31, \"wind_speed_kmh\": 41, \"conditi...", "cc": null, "timestamp": "2026-02-22T12:00:00"}
2
5
chain
retrieve-send
L1_chain_0062
L1_chain
Get the weather in Chicago and email the report to olivia@sales.deal.
[ "get_weather", "send_email", "split_text", "round_number", "web_search" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Chicago"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "send_email", "arguments": {"to": "olivia@sales.deal", "subject": "Weather in Chicago", "body": "{\"city\": \"Chicago\", \"temperature_celsius\": -5, \"humidity_percent\": 30, \"wind_speed_kmh\": 30, \"condition\": \"snowy\", \"feels_like_celsius\": -7}"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"status": "sent", "message_id": "msg_f032ff081cd7", "to": "olivia@sales.deal", "subject": "Weather in Chicago", "body_preview": "{\"city\": \"Chicago\", \"temperature_celsius\": -5, \"humidity_percent\": 30, \"wind_speed_kmh\": 30, \"condit...", "cc": null, "timestamp": "2026-02-22T12:00:00"}
2
5
chain
retrieve-send
L1_chain_0063
L1_chain
Check Chicago's weather and send it via email to karen@legal.law.
[ "get_weather", "send_email", "list_memories", "data_filter", "next_occurrence" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Chicago"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "send_email", "arguments": {"to": "karen@legal.law", "subject": "Weather in Chicago", "body": "{\"city\": \"Chicago\", \"temperature_celsius\": -5, \"humidity_percent\": 30, \"wind_speed_kmh\": 30, \"condition\": \"snowy\", \"feels_like_celsius\": -7}"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"status": "sent", "message_id": "msg_4e1a5a4074bf", "to": "karen@legal.law", "subject": "Weather in Chicago", "body_preview": "{\"city\": \"Chicago\", \"temperature_celsius\": -5, \"humidity_percent\": 30, \"wind_speed_kmh\": 30, \"condit...", "cc": null, "timestamp": "2026-02-22T12:00:00"}
2
5
chain
retrieve-send
L1_chain_0064
L1_chain
Read the file at /data/employees.csv and summarize its contents.
[ "read_file", "summarize_text", "spell_check", "format_number", "business_days_between" ]
{"steps": [{"step_id": "step_1", "tool_name": "read_file", "arguments": {"path": "/data/employees.csv"}, "depends_on": [], "output_key": "content"}, {"step_id": "step_2", "tool_name": "summarize_text", "arguments": {"text": "name,department,salary,years\nAlice,Engineering,95000,5\nBob,Marketing,72000,3\nCharlie,Engineering,105000,8\nDiana,Sales,68000,2\nEve,Engineering,88000,4\nFrank,Marketing,76000,6\nGrace,Sales,71000,3\nHenry,Engineering,112000,10"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_length": 221, "summary": "name,department,salary,years\nAlice,Engineering,95000,5\nBob,Marketing,72000,3\nCharlie,Engineering,105000,8\nDiana,Sales,68000,2\nEve,Engineering,88000,4\nFrank,Marketing,76000,6\nGrace,Sales,71000,3\nHenry,Engineering,112000,10", "summary_length": 221, "sentences_used": 1, "compression_ratio": 1.0}
2
5
chain
read-process
L1_chain_0065
L1_chain
Search for "smart city infrastructure" and classify the results into categories.
[ "web_search", "classify_text", "text_to_number", "spell_check", "dns_lookup" ]
{"steps": [{"step_id": "step_1", "tool_name": "web_search", "arguments": {"query": "smart city infrastructure"}, "depends_on": [], "output_key": "results"}, {"step_id": "step_2", "tool_name": "classify_text", "arguments": {"text": "[{\"title\": \"Smart City Infrastructure \\u2014 News Overview\", \"url\": \"https://stackoverflow.com/article/32326\", \"snippet\": \"Latest research findings on smart city infrastructure from leading institutions and peer-reviewed publications.\", \"source\": \"News\"}, {\"title\": \"Smart City Infrastructure \\u2014 ", "categories": ["science", "technology", "politics"]}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"text_preview": "[{\"title\": \"smart city infrastructure \\u2014 news overview\", \"url\": \"https://stackoverflow.com/artic...", "predicted_category": "science", "confidence": 0.0, "all_scores": {"science": 0.0, "technology": 0.0, "politics": 0.0}}
2
5
chain
retrieve-classify
L1_chain_0066
L1_chain
Search for "digital twin technology" and classify the results into categories.
[ "web_search", "classify_text", "check_url_status", "extract_entities", "detect_language" ]
{"steps": [{"step_id": "step_1", "tool_name": "web_search", "arguments": {"query": "digital twin technology"}, "depends_on": [], "output_key": "results"}, {"step_id": "step_2", "tool_name": "classify_text", "arguments": {"text": "[{\"title\": \"Digital Twin Technology \\u2014 Wikipedia Overview\", \"url\": \"https://en.wikipedia.org/article/40290\", \"snippet\": \"A comprehensive overview of digital twin technology covering key concepts, recent developments, and practical applications.\", \"source\": \"Wikipedia\"}, {\"title\": \"Digital Twin T", "categories": ["sports", "education", "environment"]}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"text_preview": "[{\"title\": \"digital twin technology \\u2014 wikipedia overview\", \"url\": \"https://en.wikipedia.org/art...", "predicted_category": "sports", "confidence": 0.0, "all_scores": {"sports": 0.0, "education": 0.0, "environment": 0.0}}
2
5
chain
retrieve-classify
L1_chain_0067
L1_chain
Find information about "autonomous vehicle regulations" and categorize it.
[ "web_search", "classify_text", "min_max", "extract_links", "spell_check" ]
{"steps": [{"step_id": "step_1", "tool_name": "web_search", "arguments": {"query": "autonomous vehicle regulations"}, "depends_on": [], "output_key": "results"}, {"step_id": "step_2", "tool_name": "classify_text", "arguments": {"text": "[{\"title\": \"Autonomous Vehicle Regulations \\u2014 News Overview\", \"url\": \"https://stackoverflow.com/article/41076\", \"snippet\": \"Latest research findings on autonomous vehicle regulations from leading institutions and peer-reviewed publications.\", \"source\": \"News\"}, {\"title\": \"Autonomous Vehicle Regu", "categories": ["science", "technology", "politics"]}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"text_preview": "[{\"title\": \"autonomous vehicle regulations \\u2014 news overview\", \"url\": \"https://stackoverflow.com/...", "predicted_category": "science", "confidence": 0.0, "all_scores": {"science": 0.0, "technology": 0.0, "politics": 0.0}}
2
5
chain
retrieve-classify
L1_chain_0068
L1_chain
Count words in "The Internet of Things connects billions of devices worldwid...", then format the number with commas.
[ "word_count", "format_number", "compress_data", "parse_html", "http_request" ]
{"steps": [{"step_id": "step_1", "tool_name": "word_count", "arguments": {"text": "The Internet of Things connects billions of devices worldwide. Smart home technology is becoming standard in new construction."}, "depends_on": [], "output_key": "wc"}, {"step_id": "step_2", "tool_name": "format_number", "arguments": {"value": 18, "format": "comma"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"formatted": "18.00"}
2
5
chain
count-format
L1_chain_0069
L1_chain
How many words are in "Urban farming and vertical agriculture are addressing food s..."? Give me the count formatted nicely.
[ "word_count", "format_number", "create_invoice", "write_file", "web_page_fetch" ]
{"steps": [{"step_id": "step_1", "tool_name": "word_count", "arguments": {"text": "Urban farming and vertical agriculture are addressing food security in cities. These innovations reduce transportation costs and carbon emissions."}, "depends_on": [], "output_key": "wc"}, {"step_id": "step_2", "tool_name": "format_number", "arguments": {"value": 19, "format": "comma"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"formatted": "19.00"}
2
5
chain
count-format
L1_chain_0070
L1_chain
Count words in "Ocean acidification threatens marine ecosystems worldwide. C...", then format the number with commas.
[ "word_count", "format_number", "create_task", "detect_language", "linear_regression" ]
{"steps": [{"step_id": "step_1", "tool_name": "word_count", "arguments": {"text": "Ocean acidification threatens marine ecosystems worldwide. Coral reefs are particularly vulnerable to changes in water chemistry."}, "depends_on": [], "output_key": "wc"}, {"step_id": "step_2", "tool_name": "format_number", "arguments": {"value": 16, "format": "comma"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"formatted": "16.00"}
2
5
chain
count-format
L1_chain_0071
L1_chain
Pull keywords from "Microplastics have been found in every environment on Earth....", then do a web search for the primary keyword.
[ "keyword_extract", "web_search", "join_texts", "get_exchange_rate", "create_contact" ]
{"steps": [{"step_id": "step_1", "tool_name": "keyword_extract", "arguments": {"text": "Microplastics have been found in every environment on Earth. Research into their health effects is intensifying."}, "depends_on": [], "output_key": "keywords"}, {"step_id": "step_2", "tool_name": "web_search", "arguments": {"query": "microplastics"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"query": "microplastics", "results": [{"title": "Microplastics — Blog Overview", "url": "https://nytimes.com/article/72137", "snippet": "microplastics: analysis of current trends, methodologies, and future directions in the field.", "source": "Blog"}, {"title": "Microplastics — Blog Analysis", "url": "https://nytimes.com/article/81047", "snippet": "microplastics: analysis of current trends, methodologies, and future directions in the field.", "source": "Blog"}, {"title": "Microplastics — Wikipedia Guide", "url": "https://en.wikipedia.org/article/33320", "snippet": "A comprehensive overview of microplastics covering key concepts, recent developments, and practical applications.", "source": "Wikipedia"}, {"title": "Microplastics — Forum Discussion", "url": "https://reuters.com/article/13829", "snippet": "Comparing different approaches to microplastics — strengths, limitations, and real-world performance.", "source": "Forum"}, {"title": "Microplastics — Blog Report", "url": "https://nature.com/article/12052", "snippet": "microplastics: analysis of current trends, methodologies, and future directions in the field.", "source": "Blog"}], "total": 5}
2
5
chain
extract-search
L1_chain_0072
L1_chain
Find the main keywords in "Biodegradable materials are replacing traditional plastics i..." and look up the most important one online.
[ "keyword_extract", "web_search", "regex_match", "retrieve_memory", "create_task" ]
{"steps": [{"step_id": "step_1", "tool_name": "keyword_extract", "arguments": {"text": "Biodegradable materials are replacing traditional plastics in packaging. Consumer demand is driving innovation in sustainable materials."}, "depends_on": [], "output_key": "keywords"}, {"step_id": "step_2", "tool_name": "web_search", "arguments": {"query": "materials"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"query": "materials", "results": [{"title": "Materials — News Overview", "url": "https://arxiv.org/article/40731", "snippet": "Latest research findings on materials from leading institutions and peer-reviewed publications.", "source": "News"}, {"title": "Materials — News Analysis", "url": "https://arxiv.org/article/92341", "snippet": "Latest research findings on materials from leading institutions and peer-reviewed publications.", "source": "News"}, {"title": "Materials — Blog Guide", "url": "https://nature.com/article/21492", "snippet": "materials: analysis of current trends, methodologies, and future directions in the field.", "source": "Blog"}, {"title": "Materials — News Discussion", "url": "https://arxiv.org/article/63961", "snippet": "Latest research findings on materials from leading institutions and peer-reviewed publications.", "source": "News"}, {"title": "Materials — Academic Report", "url": "https://bbc.com/article/68498", "snippet": "Expert guide to materials with detailed explanations, examples, and best practices for practitioners.", "source": "Academic"}], "total": 5}
2
5
chain
extract-search
L1_chain_0073
L1_chain
Pull keywords from "Biodegradable materials are replacing traditional plastics i...", then do a web search for the primary keyword.
[ "keyword_extract", "web_search", "compare_texts", "extract_entities", "base64_decode" ]
{"steps": [{"step_id": "step_1", "tool_name": "keyword_extract", "arguments": {"text": "Biodegradable materials are replacing traditional plastics in packaging. Consumer demand is driving innovation in sustainable materials."}, "depends_on": [], "output_key": "keywords"}, {"step_id": "step_2", "tool_name": "web_search", "arguments": {"query": "materials"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"query": "materials", "results": [{"title": "Materials — News Overview", "url": "https://arxiv.org/article/40731", "snippet": "Latest research findings on materials from leading institutions and peer-reviewed publications.", "source": "News"}, {"title": "Materials — News Analysis", "url": "https://arxiv.org/article/92341", "snippet": "Latest research findings on materials from leading institutions and peer-reviewed publications.", "source": "News"}, {"title": "Materials — Blog Guide", "url": "https://nature.com/article/21492", "snippet": "materials: analysis of current trends, methodologies, and future directions in the field.", "source": "Blog"}, {"title": "Materials — News Discussion", "url": "https://arxiv.org/article/63961", "snippet": "Latest research findings on materials from leading institutions and peer-reviewed publications.", "source": "News"}, {"title": "Materials — Academic Report", "url": "https://bbc.com/article/68498", "snippet": "Expert guide to materials with detailed explanations, examples, and best practices for practitioners.", "source": "Academic"}], "total": 5}
2
5
chain
extract-search
L1_chain_0074
L1_chain
Detect the language of "Bonjour, comment allez-vous aujourd'hui?..." and then translate it to English.
[ "detect_language", "translate_text", "json_extract", "generate_summary_stats", "time_since" ]
{"steps": [{"step_id": "step_1", "tool_name": "detect_language", "arguments": {"text": "Bonjour, comment allez-vous aujourd'hui?"}, "depends_on": [], "output_key": "detected"}, {"step_id": "step_2", "tool_name": "translate_text", "arguments": {"text": "Bonjour, comment allez-vous aujourd'hui?", "from_language": "und", "to_language": "en"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_text": "Bonjour, comment allez-vous aujourd'hui?", "translated_text": "Bonjour, comment allez-vous aujourd'hui? (English translation)", "from_language": "und", "to_language": "en"}
2
5
chain
detect-translate
L1_chain_0075
L1_chain
What language is "Donde esta la biblioteca mas cercana?..." in? Translate it to English.
[ "detect_language", "translate_text", "unit_convert", "web_page_fetch", "spell_check" ]
{"steps": [{"step_id": "step_1", "tool_name": "detect_language", "arguments": {"text": "Donde esta la biblioteca mas cercana?"}, "depends_on": [], "output_key": "detected"}, {"step_id": "step_2", "tool_name": "translate_text", "arguments": {"text": "Donde esta la biblioteca mas cercana?", "from_language": "fr", "to_language": "en"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_text": "Donde esta la biblioteca mas cercana?", "translated_text": "Donde esta la biblioteca mas cercana? (English translation)", "from_language": "fr", "to_language": "en"}
2
5
chain
detect-translate
L1_chain_0076
L1_chain
What language is "Eu gostaria de reservar uma mesa para dois...." in? Translate it to English.
[ "detect_language", "translate_text", "set_reminder", "correlation", "percentile" ]
{"steps": [{"step_id": "step_1", "tool_name": "detect_language", "arguments": {"text": "Eu gostaria de reservar uma mesa para dois."}, "depends_on": [], "output_key": "detected"}, {"step_id": "step_2", "tool_name": "translate_text", "arguments": {"text": "Eu gostaria de reservar uma mesa para dois.", "from_language": "es", "to_language": "en"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"original_text": "Eu gostaria de reservar uma mesa para dois.", "translated_text": "Eu gostaria de reservar uma mesa para dois. (English translation)", "from_language": "es", "to_language": "en"}
2
5
chain
detect-translate
L1_chain_0077
L1_chain
Find the % change from 383.0 to 195.77 and present it in a formatted style.
[ "percentage_change", "format_number", "web_search", "data_aggregate", "create_contact" ]
{"steps": [{"step_id": "step_1", "tool_name": "percentage_change", "arguments": {"old_value": 383.0, "new_value": 195.77}, "depends_on": [], "output_key": "pct"}, {"step_id": "step_2", "tool_name": "format_number", "arguments": {"value": -48.8851, "format": "percent"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"formatted": "-4888.51%"}
2
5
chain
compute-format
L1_chain_0078
L1_chain
Compute the percentage difference from 19.5 to 274.32 and display it formatted.
[ "percentage_change", "format_number", "write_file", "json_extract", "data_sort" ]
{"steps": [{"step_id": "step_1", "tool_name": "percentage_change", "arguments": {"old_value": 19.5, "new_value": 274.32}, "depends_on": [], "output_key": "pct"}, {"step_id": "step_2", "tool_name": "format_number", "arguments": {"value": 1306.7692, "format": "percent"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"formatted": "130676.92%"}
2
5
chain
compute-format
L1_chain_0079
L1_chain
What is the percentage change from 392.8 to 329.36? Format the result as a percentage.
[ "percentage_change", "format_number", "merge_data", "get_directions", "create_notification" ]
{"steps": [{"step_id": "step_1", "tool_name": "percentage_change", "arguments": {"old_value": 392.8, "new_value": 329.36}, "depends_on": [], "output_key": "pct"}, {"step_id": "step_2", "tool_name": "format_number", "arguments": {"value": -16.1507, "format": "percent"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"formatted": "-1615.07%"}
2
5
chain
compute-format
L1_chain_0080
L1_chain
Find the numbers in "Budget allocated $5.2 million for research, $3.1 million for..." and run a statistical analysis.
[ "extract_numbers", "statistical_analysis", "merge_data", "time_since", "create_invoice" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_numbers", "arguments": {"text": "Budget allocated $5.2 million for research, $3.1 million for development, and $1.7 million for marketing."}, "depends_on": [], "output_key": "extracted"}, {"step_id": "step_2", "tool_name": "statistical_analysis", "arguments": {"numbers": [5.2, 3.1, 1.7], "operation": "summary"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"count": 3, "mean": 3.3333, "median": 3.1, "min": 1.7, "max": 5.2, "sum": 10.0, "stdev": 1.7616}
2
5
chain
extract-analyze
L1_chain_0081
L1_chain
Find the numbers in "Population grew from 8.3 million in 2010 to 9.1 million in 2..." and run a statistical analysis.
[ "extract_numbers", "statistical_analysis", "create_contact", "slugify", "database_query" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_numbers", "arguments": {"text": "Population grew from 8.3 million in 2010 to 9.1 million in 2020, an increase of 800,000."}, "depends_on": [], "output_key": "extracted"}, {"step_id": "step_2", "tool_name": "statistical_analysis", "arguments": {"numbers": [8.3, 201, 0, 9.1, 202, 0, 800000], "operation": "summary"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"count": 7, "mean": 114345.7714, "median": 9.1, "min": 0.0, "max": 800000.0, "sum": 800420.4, "stdev": 302345.11}
2
5
chain
extract-analyze
L1_chain_0082
L1_chain
Find the numbers in "Inflation hit 6.2 percent in 2022, up from 1.4 percent the p..." and run a statistical analysis.
[ "extract_numbers", "statistical_analysis", "convert_timezone", "parse_html", "split_text" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_numbers", "arguments": {"text": "Inflation hit 6.2 percent in 2022, up from 1.4 percent the previous year."}, "depends_on": [], "output_key": "extracted"}, {"step_id": "step_2", "tool_name": "statistical_analysis", "arguments": {"numbers": [6.2, 202, 2, 1.4], "operation": "summary"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"count": 4, "mean": 52.9, "median": 4.1, "min": 1.4, "max": 202.0, "sum": 211.6, "stdev": 99.4229}
2
5
chain
extract-analyze
L1_chain_0083
L1_chain
What day of the week is "Jan 1, 2027"? Parse the date first.
[ "parse_date", "get_weekday", "normalize_data", "add_duration", "classify_text" ]
{"steps": [{"step_id": "step_1", "tool_name": "parse_date", "arguments": {"date_string": "Jan 1, 2027"}, "depends_on": [], "output_key": "parsed"}, {"step_id": "step_2", "tool_name": "get_weekday", "arguments": {"date": "2027-01-01"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"weekday": "Friday", "day_number": 4}
2
5
chain
parse-lookup
L1_chain_0084
L1_chain
Convert "March 15, 2026" to a standard date, then find out the weekday.
[ "parse_date", "get_weekday", "sentiment_analysis", "word_count", "spell_check" ]
{"steps": [{"step_id": "step_1", "tool_name": "parse_date", "arguments": {"date_string": "March 15, 2026"}, "depends_on": [], "output_key": "parsed"}, {"step_id": "step_2", "tool_name": "get_weekday", "arguments": {"date": "2026-03-15"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"weekday": "Sunday", "day_number": 6}
2
5
chain
parse-lookup
L1_chain_0085
L1_chain
What day of the week is "05/05/2026"? Parse the date first.
[ "parse_date", "get_weekday", "execute_python", "compare_texts", "format_number" ]
{"steps": [{"step_id": "step_1", "tool_name": "parse_date", "arguments": {"date_string": "05/05/2026"}, "depends_on": [], "output_key": "parsed"}, {"step_id": "step_2", "tool_name": "get_weekday", "arguments": {"date": "2026-05-05"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"weekday": "Tuesday", "day_number": 1}
2
5
chain
parse-lookup
L1_chain_0086
L1_chain
Open /data/employees.csv and pull out the most important keywords.
[ "read_file", "keyword_extract", "add_duration", "slugify", "dns_lookup" ]
{"steps": [{"step_id": "step_1", "tool_name": "read_file", "arguments": {"path": "/data/employees.csv"}, "depends_on": [], "output_key": "content"}, {"step_id": "step_2", "tool_name": "keyword_extract", "arguments": {"text": "name,department,salary,years\nAlice,Engineering,95000,5\nBob,Marketing,72000,3\nCharlie,Engineering,105000,8\nDiana,Sales,68000,2\nEve,Engineering,88000,4\nFrank,Marketing,76000,6\nGrace,Sales,71000,3\nHenry,Engineering,112000,10"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"keywords": ["engineering", "marketing", "sales", "name", "department"], "scores": [1.0, 0.5, 0.5, 0.25, 0.25]}
2
5
chain
read-extract
L1_chain_0087
L1_chain
Read the file at /data/config.json and extract the keywords from its contents.
[ "read_file", "keyword_extract", "encode_url", "extract_entities", "rss_feed_parse" ]
{"steps": [{"step_id": "step_1", "tool_name": "read_file", "arguments": {"path": "/data/config.json"}, "depends_on": [], "output_key": "content"}, {"step_id": "step_2", "tool_name": "keyword_extract", "arguments": {"text": "{\"app_name\": \"DataPipeline\", \"version\": \"2.1.0\", \"max_workers\": 4, \"timeout_seconds\": 30, \"features\": {\"caching\": true, \"logging\": true, \"metrics\": false}}"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"keywords": ["true", "datapipeline", "version", "features", "caching"], "scores": [1.0, 0.5, 0.5, 0.5, 0.5]}
2
5
chain
read-extract
L1_chain_0088
L1_chain
Read /data/report.txt, then identify the key topics and keywords.
[ "read_file", "keyword_extract", "retrieve_memory", "generate_summary_stats", "base64_decode" ]
{"steps": [{"step_id": "step_1", "tool_name": "read_file", "arguments": {"path": "/data/report.txt"}, "depends_on": [], "output_key": "content"}, {"step_id": "step_2", "tool_name": "keyword_extract", "arguments": {"text": "Q1 2026 Sales Report\n\nTotal revenue: $2.4M\nGrowth: 15% YoY\nTop product: Enterprise Plan ($1.2M)\nNew customers: 340\nChurn rate: 2.1%\n\nKey highlights:\n- Enterprise segment grew 28%\n- APAC region exceeded targets by 12%\n- Customer satisfaction score: 4.6/5.0"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"keywords": ["enterprise", "sales", "report", "total", "revenue"], "scores": [1.0, 0.5, 0.5, 0.5, 0.5]}
2
5
chain
read-extract
L1_chain_0089
L1_chain
Read the file at /data/config.json and count how many words it contains.
[ "read_file", "word_count", "log_event", "url_parse", "percentile" ]
{"steps": [{"step_id": "step_1", "tool_name": "read_file", "arguments": {"path": "/data/config.json"}, "depends_on": [], "output_key": "content"}, {"step_id": "step_2", "tool_name": "word_count", "arguments": {"text": "{\"app_name\": \"DataPipeline\", \"version\": \"2.1.0\", \"max_workers\": 4, \"timeout_seconds\": 30, \"features\": {\"caching\": true, \"logging\": true, \"metrics\": false}}"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"words": 15, "characters": 155, "sentences": 2}
2
5
chain
read-count
L1_chain_0090
L1_chain
How many words are in the file at /data/report.txt?
[ "read_file", "word_count", "generate_report", "add_duration", "generate_summary_stats" ]
{"steps": [{"step_id": "step_1", "tool_name": "read_file", "arguments": {"path": "/data/report.txt"}, "depends_on": [], "output_key": "content"}, {"step_id": "step_2", "tool_name": "word_count", "arguments": {"text": "Q1 2026 Sales Report\n\nTotal revenue: $2.4M\nGrowth: 15% YoY\nTop product: Enterprise Plan ($1.2M)\nNew customers: 340\nChurn rate: 2.1%\n\nKey highlights:\n- Enterprise segment grew 28%\n- APAC region exceeded targets by 12%\n- Customer satisfaction score: 4.6/5.0"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"words": 40, "characters": 255, "sentences": 5}
2
5
chain
read-count
L1_chain_0091
L1_chain
How many words are in the file at /data/config.json?
[ "read_file", "word_count", "truncate_text", "business_days_between", "generate_summary_stats" ]
{"steps": [{"step_id": "step_1", "tool_name": "read_file", "arguments": {"path": "/data/config.json"}, "depends_on": [], "output_key": "content"}, {"step_id": "step_2", "tool_name": "word_count", "arguments": {"text": "{\"app_name\": \"DataPipeline\", \"version\": \"2.1.0\", \"max_workers\": 4, \"timeout_seconds\": 30, \"features\": {\"caching\": true, \"logging\": true, \"metrics\": false}}"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"words": 15, "characters": 155, "sentences": 2}
2
5
chain
read-count
L1_chain_0092
L1_chain
Look up the weather in Beijing, then log this as an info event.
[ "get_weather", "log_event", "calculate_date_diff", "generate_url", "retrieve_memory" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Beijing"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "log_event", "arguments": {"event_type": "weather_check", "message": "Weather in Beijing: cloudy, 19°C", "severity": "info"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"log_id": "log_065f87e80706", "event_type": "weather_check", "message": "Weather in Beijing: cloudy, 19°C", "severity": "info", "timestamp": "2026-02-22T12:00:00"}
2
5
chain
retrieve-log
L1_chain_0093
L1_chain
Fetch Manila's weather and record it in the event log.
[ "get_weather", "log_event", "set_reminder", "generate_summary_stats", "case_convert" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Manila"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "log_event", "arguments": {"event_type": "weather_check", "message": "Weather in Manila: foggy, 26°C", "severity": "info"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"log_id": "log_fe8e21a90258", "event_type": "weather_check", "message": "Weather in Manila: foggy, 26°C", "severity": "info", "timestamp": "2026-02-22T12:00:00"}
2
5
chain
retrieve-log
L1_chain_0094
L1_chain
Look up the weather in Lisbon, then log this as an info event.
[ "get_weather", "log_event", "correlation", "web_search", "data_filter" ]
{"steps": [{"step_id": "step_1", "tool_name": "get_weather", "arguments": {"city": "Lisbon"}, "depends_on": [], "output_key": "weather"}, {"step_id": "step_2", "tool_name": "log_event", "arguments": {"event_type": "weather_check", "message": "Weather in Lisbon: snowy, 16°C", "severity": "info"}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"log_id": "log_499c52eb4b19", "event_type": "weather_check", "message": "Weather in Lisbon: snowy, 16°C", "severity": "info", "timestamp": "2026-02-22T12:00:00"}
2
5
chain
retrieve-log
L1_chain_0095
L1_chain
Evaluate 295 * 59 and give me the answer rounded to 3 places.
[ "calculator", "round_number", "standard_deviation", "create_task", "log_event" ]
{"steps": [{"step_id": "step_1", "tool_name": "calculator", "arguments": {"expression": "295 * 59"}, "depends_on": [], "output_key": "calc"}, {"step_id": "step_2", "tool_name": "round_number", "arguments": {"value": 17405.0, "decimals": 3}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"rounded": 17405.0}
2
5
chain
compute-round
L1_chain_0096
L1_chain
Compute 881 * 43, then round the answer to 3 decimals.
[ "calculator", "round_number", "generate_image", "extract_domain", "web_page_fetch" ]
{"steps": [{"step_id": "step_1", "tool_name": "calculator", "arguments": {"expression": "881 * 43"}, "depends_on": [], "output_key": "calc"}, {"step_id": "step_2", "tool_name": "round_number", "arguments": {"value": 37883.0, "decimals": 3}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"rounded": 37883.0}
2
5
chain
compute-round
L1_chain_0097
L1_chain
Compute 358 + 37, then round the answer to 0 decimals.
[ "calculator", "round_number", "data_filter", "merge_data", "join_texts" ]
{"steps": [{"step_id": "step_1", "tool_name": "calculator", "arguments": {"expression": "358 + 37"}, "depends_on": [], "output_key": "calc"}, {"step_id": "step_2", "tool_name": "round_number", "arguments": {"value": 395.0, "decimals": 0}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"rounded": 395.0}
2
5
chain
compute-round
L1_chain_0098
L1_chain
Get the numeric values in "The garden covers 2.5 acres with over 300 species of plants ..." and compute their range.
[ "extract_numbers", "min_max", "summarize_text", "check_url_status", "transform_format" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_numbers", "arguments": {"text": "The garden covers 2.5 acres with over 300 species of plants and 15 fountains."}, "depends_on": [], "output_key": "extracted"}, {"step_id": "step_2", "tool_name": "min_max", "arguments": {"numbers": [2.5, 300, 15]}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"min": 2.5, "max": 300.0, "range": 297.5}
2
5
chain
extract-aggregate
L1_chain_0099
L1_chain
Get the numeric values in "The satellite orbits at 35,786 km altitude completing 1 revo..." and compute their range.
[ "extract_numbers", "min_max", "create_calendar_event", "encode_url", "create_notification" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_numbers", "arguments": {"text": "The satellite orbits at 35,786 km altitude completing 1 revolution every 24 hours."}, "depends_on": [], "output_key": "extracted"}, {"step_id": "step_2", "tool_name": "min_max", "arguments": {"numbers": [35786, 1, 24]}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"min": 1.0, "max": 35786.0, "range": 35785.0}
2
5
chain
extract-aggregate
L1_chain_0100
L1_chain
Pull out the numbers from "The recipe calls for 2.5 cups of flour, 3 eggs, and 175 gram...", then identify the smallest and largest.
[ "extract_numbers", "min_max", "write_file", "text_similarity", "format_number" ]
{"steps": [{"step_id": "step_1", "tool_name": "extract_numbers", "arguments": {"text": "The recipe calls for 2.5 cups of flour, 3 eggs, and 175 grams of sugar."}, "depends_on": [], "output_key": "extracted"}, {"step_id": "step_2", "tool_name": "min_max", "arguments": {"numbers": [2.5, 3, 175]}, "depends_on": ["step_1"], "output_key": null}], "final_answer_source": "step_2"}
{"min": 2.5, "max": 175.0, "range": 172.5}
2
5
chain
extract-aggregate
End of preview. Expand in Data Studio

CompToolBench: Measuring Compositional Tool-Use Generalization in LLMs

Dataset Summary

CompToolBench is a benchmark for evaluating how well large language models generalize from simple, single-tool calls to complex, multi-step compositional tool use. It contains 200 tasks spanning four composition levels of increasing structural complexity, built on top of 106 deterministic tool simulators covering 9 functional categories.

The key insight behind CompToolBench is the composition gap: models that can reliably call individual tools often fail dramatically when those same tools must be composed into chains, parallel fan-outs, or directed acyclic graphs (DAGs). CompToolBench quantifies this gap with fine-grained diagnostic metrics.

Key Features

  • 4 composition levels: single calls (L0), sequential chains (L1), parallel fan-outs (L2), and full DAGs with branching + merging (L3)
  • 106 deterministic tool simulators: no external API dependencies, fully reproducible
  • Fine-grained scoring: tool selection accuracy, argument accuracy, data-flow correctness, completion rate
  • 18-model leaderboard: spanning cloud APIs (Mistral, Cohere, Groq, Cerebras, OpenRouter) and local models (Ollama)
  • Composition gap metric: directly measures how much accuracy degrades as structural complexity increases

Dataset Structure

Each example in the dataset contains the following fields:

Field Type Description
task_id string Unique identifier (e.g., L0_node_0001, L3_dag_0153)
level string Composition level: L0_node, L1_chain, L2_parallel, or L3_dag
prompt string Natural language instruction given to the model
available_tools list[string] Tool names provided to the model (includes distractors)
expected_trace object Ground-truth execution plan with steps, dependencies, and arguments
expected_final_answer string JSON-serialized expected output
num_steps int Number of tool calls in the expected trace
num_tools_offered int Number of tools offered (correct + distractors)
category string Functional category of the task
pattern string Composition pattern (e.g., retrieve-transform, fan-out-compare)

Expected Trace Structure

Each step in expected_trace.steps contains:

  • step_id: Step identifier (e.g., step_1)
  • tool_name: Which tool to call
  • arguments: JSON-serialized expected arguments
  • depends_on: List of step IDs this step depends on (defines the DAG structure)
  • output_key: Variable name for the step's output (used by downstream steps)

Composition Levels

Level Name Description Tasks Avg Steps Avg Tools Offered
L0 Single Node One tool call, no composition 48 1.0 4.0
L1 Chain Sequential pipeline (A -> B) 64 2.0 5.0
L2 Parallel Independent fan-out (A || B || C) 40 2.8 4.2
L3 DAG Full directed acyclic graph with branching and merging 48 4.4 6.6

Task Categories

Tasks cover 9 functional categories: chain, communication, computation, dag, external_services, information_retrieval, parallel, text_processing, and time_scheduling.

Composition Patterns

Over 40 distinct composition patterns are represented, including retrieve-transform, fan-out-compare, chain-fanout-merge-chain, parallel-merge-chain, true-dag-parallel-reads-merge, and many more. See the paper for full details.

Leaderboard

Results from evaluating 18 models (10 cloud, 8 local) on all 200 tasks. Models are ranked by overall accuracy. All models achieve 100% tool selection accuracy (when they issue a call, they name the correct tool).

Cloud Models

Model Provider L0 L1 L2 L3 Overall Delta
Llama 3.1 8B Groq 27.1 75.8 87.1 76.0 66.4 -48.9
Command A Cohere 45.8 62.7 87.8 40.8 58.4 5.1
Mistral Small Mistral 45.8 59.7 87.6 40.9 57.5 4.9
Command R+ Cohere 43.8 57.5 88.0 40.3 56.2 3.4
Llama 3.1 8B Cerebras 31.2 66.1 81.2 46.4 56.0 -15.1
Mistral Large Mistral 39.6 59.5 87.9 38.5 55.4 1.1
Mistral Medium Mistral 43.8 57.5 87.9 36.3 55.2 7.4
Gemini 2.0 Flash OpenRouter 39.6 52.4 85.7 39.0 52.8 0.6
GPT-OSS 120B Cerebras 45.8 56.3 56.1 29.0 47.2 16.8
Llama 4 Scout 17B Groq 37.5 49.6 55.8 7.0 37.7 30.5

Local Models (Ollama)

Model Provider L0 L1 L2 L3 Overall Delta
Granite4 3B Ollama 45.8 57.3 56.1 30.2 47.8 15.6
Granite4 1B Ollama 41.7 56.3 55.9 29.9 46.4 11.8
Mistral 7B Ollama 43.8 57.7 49.2 30.5 46.1 13.3
Llama 3.1 8B Ollama 39.6 56.7 56.1 29.5 45.9 10.1
Mistral Nemo 12B Ollama 37.5 58.4 51.0 31.8 45.5 5.7
Qwen 2.5 7B Ollama 39.6 56.7 53.8 25.8 44.6 13.8
Mistral Small 24B Ollama 37.5 51.1 47.7 22.6 40.3 14.9
Qwen3 8B Ollama 35.4 52.0 36.9 21.8 37.7 13.7

Aggregate Statistics

Segment L0 L1 L2 L3 Overall Delta
All models avg. 40.0 58.0 67.3 34.2 49.8 5.8
Cloud avg. 40.0 59.7 80.5 39.4 54.3 0.6
Local avg. 40.1 55.8 50.8 27.8 44.3 12.3

Delta = L0 accuracy minus L3 accuracy (positive means degradation at higher composition levels). Models marked with a dagger in the paper exhibit a Selection Gap, where L0 accuracy is lower than the average of L1-L3.

Usage

Loading the Dataset

from datasets import load_dataset

dataset = load_dataset("mdarahmanxAI/comptoolbench", split="test")

# Browse tasks by composition level
l3_tasks = dataset.filter(lambda x: x["level"] == "L3_dag")
print(f"L3 DAG tasks: {len(l3_tasks)}")
print(l3_tasks[0]["prompt"])

Evaluating a Model

CompToolBench evaluates models by comparing their tool-call traces against the expected trace. The evaluation harness is available in the GitHub repository.

import json

for task in dataset:
    # 1. Build the tool-use prompt from task["prompt"] and task["available_tools"]
    # 2. Send to your model with the tool schemas
    # 3. Compare the model's tool calls against:
    trace = json.loads(task["expected_trace"])
    answer = json.loads(task["expected_final_answer"])

    # Scoring dimensions:
    #   - Tool selection: did the model call the right tools?
    #   - Argument accuracy: were the arguments correct?
    #   - Data flow: did outputs flow correctly between steps?
    #   - Completion: did all required steps execute?

Scoring Metrics

Metric Description
Overall Accuracy Weighted combination of all sub-metrics
Tool Selection Whether the model called the correct tool names
Argument Accuracy Whether arguments matched expected values
Data Flow Accuracy Whether inter-step data dependencies were satisfied
Completion Rate Fraction of expected steps that were executed
Composition Gap L0 accuracy minus Lk accuracy (measures degradation)

Citation

If you use CompToolBench in your research, please cite:

@article{rahmaan2026comptoolbench,
  title={CompToolBench: Measuring Compositional Tool-Use Generalization in Large Language Models},
  author={Rahmaan, Rony},
  journal={arXiv preprint},
  year={2026}
}

License

This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. You are free to share and adapt the dataset for any purpose, provided you give appropriate credit.

Links

Downloads last month
11