Toolathlon

course-schedule

Pick this student's classes for the semester out of a timetable and a master exam table, then write the exams they will sit to exam_schedule.jsonl, sorted by when they start. Who the student is — their id, their department, their English score — is in Memory.

Open in Gym

Task Prompt

My personal information is stored in Memory. The workspace contains my course schedule and a master exam table (in Excel format). The courses I need to take this semester are in list.txt.
Please help me choose the right courses. I can only take one of each course. I have club activities in the evening, so I'll only take daytime courses. Also, my favorite teacher is Yulian(郁莲). if she teaches any required courses, please make sure to take her.
Based on my personal information, please first help me find the right specific courses. Then, help me find the exams I need to take this semester. Save the exam information in the workspace in a file named exam_schedule.jsonl. The format should refer to `format.md`.
All information must match the Excel file. If a course is not required this semester(Course exemption), omit it. Finally, sort the objects in the JSONL file by exam time.

VFS.md

---
name: workspace-filesystem
description: The virtual filesystem layout of this workspace.
---

# Virtual Filesystem

You work in a terminal over a virtual filesystem. Read files with shell
commands; service operations are listed in SKILL.md.

| Path | Description |
|---|---|
| `/workspace` | Agent workspace; current directory and home (`~`). |
| `/tmp` | Temporary files. |
| `/dev` | Device files: `null` and `zero` (mounted automatically). |

name	workspace-filesystem
description	The virtual filesystem layout of this workspace.

Virtual Filesystem

You work in a terminal over a virtual filesystem. Read files with shell commands; service operations are listed in SKILL.md.

Path	Description
`/workspace`	Agent workspace; current directory and home (`~`).
`/tmp`	Temporary files.
`/dev`	Device files: `null` and `zero` (mounted automatically).

SKILL.md

---
name: workspace-commands
description: The shell commands, task tools, and python3 interpreter available in this workspace.
---

# Available commands

Standard Unix commands, among them:

    cat  head  tail  ls  find  grep  wc  sort  uniq  cut  tr  sed  awk
    diff  tee  xargs  jq  unzip  curl

with pipes, `>` redirection, `$(...)` substitution and `$'...'`.

The task CLIs below are registered with this shell rather than installed
as files; do not inspect them with `which`, `file` or `/usr/bin`.

## pdf-tools

Use `pdf-tools <command> '<JSON arguments>'`. Run
`pdf-tools <command> --help` to get more information.

| Command | Arguments | Description |
|---|---|---|
| `pdf-tools read-pdf-pages` | `{pdf_file_path: str, start_page?: int = 1, end_page?: int = 1}` | Read content from PDF file for specified page range. Supports both local file paths and URLs. For URLs, the PDF will be downloaded to a temporary directory and cached for future use. Note: Avoid reading too many pages at once (recommended: <50 pages) to prevent errors. Args: pdf_file_path: Path to the PDF file or URL to PDF start_page: Starting page number (default: 1) end_page: Ending page number (default: 1) Returns: Extracted text content from the specified pages |
| `pdf-tools get-pdf-info` | `{pdf_file_path: str}` | Get basic information about a PDF file including page count. Supports both local file paths and URLs. For URLs, the PDF will be downloaded to a temporary directory and cached for future use. Args: pdf_file_path: Path to the PDF file or URL to PDF Returns: Basic information about the PDF file |
| `pdf-tools merge-pdfs` | `{pdf_paths: list[str], output_path: str}` | Merge multiple PDF files into one. Supports both local file paths and URLs. URLs will be downloaded to temporary files before merging. Mixed local and URL paths are supported. Args: pdf_paths: List of paths to PDF files to merge (in order) - can include URLs output_path: Path where the merged PDF will be saved (must be local path) Returns: Success message with merge details or error message |
| `pdf-tools extract-pdf-pages` | `{source_path: str, page_numbers: list[int], output_path: str}` | Extract specific pages from a PDF and create a new PDF. Supports URLs for source PDF. The source PDF will be downloaded to a temporary directory if it's a URL. Output path must be a local file path. Args: source_path: Path to the source PDF file or URL to PDF page_numbers: List of page numbers to extract (1-indexed) output_path: Path where the new PDF will be saved (must be local path) Returns: Success message with extraction details or error message |
| `pdf-tools search-pdf-content` | `{pdf_file_path: str, pattern: str, page_size?: int = 10}` | Search for regex pattern in PDF content and return paginated results. Supports both local file paths and URLs. For URLs, the PDF will be downloaded to a temporary directory and cached for future use. Args: pdf_file_path: Path to the PDF file or URL to PDF pattern: Regular expression pattern to search for page_size: Number of results per page (10-50, default: 10) Returns: Search results with UUID for pagination, or error message |
| `pdf-tools search-pdf-next-page` | `{search_id: str}` | Get next page of search results. Args: search_id: Search session ID from previous search Returns: Next page of search results or error message |
| `pdf-tools search-pdf-prev-page` | `{search_id: str}` | Get previous page of search results. Args: search_id: Search session ID from previous search Returns: Previous page of search results or error message |
| `pdf-tools search-pdf-go-page` | `{search_id: str, page_number: int}` | Go to specific page of search results. Args: search_id: Search session ID from previous search page_number: Page number to go to (1-indexed) Returns: Specified page of search results or error message |
| `pdf-tools search-pdf-info` | `{search_id: str}` | Get information about a search session. Args: search_id: Search session ID from previous search Returns: Information about the search session |

## excel

Use `excel <command> '<JSON arguments>'`. Run `excel <command> --help` to
get more information.

| Command | Arguments | Description |
|---|---|---|
| `excel apply-formula` | `{filepath: str, sheet_name: str, cell: str, formula: str}` | Apply Excel formula to cell. Excel formula will write to cell with verification. |
| `excel validate-formula-syntax` | `{filepath: str, sheet_name: str, cell: str, formula: str}` | Validate Excel formula syntax without applying it. |
| `excel format-range` | `{filepath: str, sheet_name: str, start_cell: str, end_cell?: str = null, bold?: bool = false, italic?: bool = false, underline?: bool = false, font_size?: int = null, font_color?: str = null, bg_color?: str = null, border_style?: str = null, border_color?: str = null, number_format?: str = null, alignment?: str = null, wrap_text?: bool = false, merge_cells?: bool = false, protection?: dict = null, conditional_format?: dict = null}` | Apply formatting to a range of cells. |
| `excel read-data-from-excel` | `{filepath: str, sheet_name: str, start_cell?: str = "A1", end_cell?: str = null, preview_only?: bool = false}` | Read data from Excel worksheet with cell metadata including validation rules. Args: filepath: Path to Excel file sheet_name: Name of worksheet start_cell: Starting cell (default A1) end_cell: Ending cell (optional, auto-expands if not provided) preview_only: Whether to return preview only Returns: JSON string containing structured cell data with validation metadata. Each cell includes: address, value, row, column, and validation info (if any). |
| `excel write-data-to-excel` | `{filepath: str, sheet_name: str, data: list[list], start_cell?: str = "A1"}` | Write data to Excel worksheet. Excel formula will write to cell without any verification. PARAMETERS: filepath: Path to Excel file sheet_name: Name of worksheet to write to data: List of lists containing data to write to the worksheet, sublists are assumed to be rows start_cell: Cell to start writing to, default is "A1" |
| `excel create-workbook` | `{filepath: str}` | Create new Excel workbook. |
| `excel create-worksheet` | `{filepath: str, sheet_name: str}` | Create new worksheet in workbook. |
| `excel create-chart` | `{filepath: str, sheet_name: str, data_range: str, chart_type: str, target_cell: str, title?: str = "", x_axis?: str = "", y_axis?: str = ""}` | Create chart in worksheet. |
| `excel create-pivot-table` | `{filepath: str, sheet_name: str, data_range: str, rows: list[str], values: list[str], columns?: list[str] = null, agg_func?: str = "mean"}` | Create pivot table in worksheet. |
| `excel create-table` | `{filepath: str, sheet_name: str, data_range: str, table_name?: str = null, table_style?: str = "TableStyleMedium9"}` | Creates a native Excel table from a specified range of data. |
| `excel copy-worksheet` | `{filepath: str, source_sheet: str, target_sheet: str}` | Copy worksheet within workbook. |
| `excel delete-worksheet` | `{filepath: str, sheet_name: str}` | Delete worksheet from workbook. |
| `excel rename-worksheet` | `{filepath: str, old_name: str, new_name: str}` | Rename worksheet in workbook. |
| `excel get-workbook-metadata` | `{filepath: str, include_ranges?: bool = false}` | Get metadata about workbook including sheets, ranges, etc. |
| `excel merge-cells` | `{filepath: str, sheet_name: str, start_cell: str, end_cell: str}` | Merge a range of cells. |
| `excel unmerge-cells` | `{filepath: str, sheet_name: str, start_cell: str, end_cell: str}` | Unmerge a range of cells. |
| `excel copy-range` | `{filepath: str, sheet_name: str, source_start: str, source_end: str, target_start: str, target_sheet?: str = null}` | Copy a range of cells to another location. |
| `excel delete-range` | `{filepath: str, sheet_name: str, start_cell: str, end_cell: str, shift_direction?: str = "up"}` | Delete a range of cells and shift remaining cells. |
| `excel validate-excel-range` | `{filepath: str, sheet_name: str, start_cell: str, end_cell?: str = null}` | Validate if a range exists and is properly formatted. |
| `excel get-data-validation-info` | `{filepath: str, sheet_name: str}` | Get all data validation rules in a worksheet. This tool helps identify which cell ranges have validation rules and what types of validation are applied. Args: filepath: Path to Excel file sheet_name: Name of worksheet Returns: JSON string containing all validation rules in the worksheet |

## memory

Use `memory <command> '<JSON arguments>'`. Run `memory <command> --help` to
get more information.

| Command | Arguments | Description |
|---|---|---|
| `memory create-entities` | `{entities: list[{name: str, entityType: str, observations: list[str]}]}` | Create multiple new entities in the knowledge graph — `entities.name`: The name of the entity; `entities.entityType`: The type of the entity; `entities.observations`: An array of observation contents associated with the entity |
| `memory create-relations` | `{relations: list[{from: str, to: str, relationType: str}]}` | Create multiple new relations between entities in the knowledge graph. Relations should be in active voice — `relations.from`: The name of the entity where the relation starts; `relations.to`: The name of the entity where the relation ends; `relations.relationType`: The type of the relation |
| `memory add-observations` | `{observations: list[{entityName: str, contents: list[str]}]}` | Add new observations to existing entities in the knowledge graph — `observations.entityName`: The name of the entity to add the observations to; `observations.contents`: An array of observation contents to add |
| `memory delete-entities` | `{entityNames: list[str]}` | Delete multiple entities and their associated relations from the knowledge graph — `entityNames`: An array of entity names to delete |
| `memory delete-observations` | `{deletions: list[{entityName: str, observations: list[str]}]}` | Delete specific observations from entities in the knowledge graph — `deletions.entityName`: The name of the entity containing the observations; `deletions.observations`: An array of observations to delete |
| `memory delete-relations` | `{relations: list[{from: str, to: str, relationType: str}]}` | Delete multiple relations from the knowledge graph — `relations`: An array of relations to delete; `relations.from`: The name of the entity where the relation starts; `relations.to`: The name of the entity where the relation ends; `relations.relationType`: The type of the relation |
| `memory read-graph` | `{}` | Read the entire knowledge graph |
| `memory search-nodes` | `{query: str}` | Search for nodes in the knowledge graph based on a query — `query`: The search query to match against entity names, types, and observation content |
| `memory open-nodes` | `{names: list[str]}` | Open specific nodes in the knowledge graph by their names — `names`: An array of entity names to retrieve |

## fetch

Use `fetch <command> '<JSON arguments>'`. Run `fetch <command> --help` to
get more information.

| Command | Arguments | Description |
|---|---|---|
| `fetch html` | `{url: str, headers?: dict}` | Fetch a website and return the content as HTML — `url`: URL of the website to fetch; `headers`: Optional headers to include in the request |
| `fetch markdown` | `{url: str, headers?: dict}` | Fetch a website and return the content as Markdown — `url`: URL of the website to fetch; `headers`: Optional headers to include in the request |
| `fetch txt` | `{url: str, headers?: dict}` | Fetch a website, return the content as plain text (no HTML) — `url`: URL of the website to fetch; `headers`: Optional headers to include in the request |
| `fetch json` | `{url: str, headers?: dict}` | Fetch a JSON file from a URL — `url`: URL of the JSON to fetch; `headers`: Optional headers to include in the request |

## web-search

Use `web-search <command> '<JSON arguments>'`. Run
`web-search <command> --help` to get more information.

| Command | Arguments | Description |
|---|---|---|
| `web-search web-search` | `{query: str, num_results?: int = 10}` | Search the web using Google Serper API with concurrency control and retry mechanisms. Supports various Google search operators. — `query`: Search query with optional Google search operators.; `num_results`: Number of results to return, default 10, max 50 |

## python3

Execute Python code directly under the agent workspace and return stdout
and stderr. Code can be pasted into a `.py` file and run with `python3`.

Python cannot launch subprocesses or make direct network requests. Run the
documented service commands directly in the shell. Use Python for local
file processing and generating JSON arguments.

name	workspace-commands
description	The shell commands, task tools, and python3 interpreter available in this workspace.

Available commands

Standard Unix commands, among them:

cat  head  tail  ls  find  grep  wc  sort  uniq  cut  tr  sed  awk
diff  tee  xargs  jq  unzip  curl

with pipes, > redirection, $(...) substitution and $'...'.

The task CLIs below are registered with this shell rather than installed as files; do not inspect them with which, file or /usr/bin.

pdf-tools

Use pdf-tools <command> '<JSON arguments>'. Run pdf-tools <command> --help to get more information.

Command	Arguments	Description
`pdf-tools read-pdf-pages`	`{pdf_file_path: str, start_page?: int = 1, end_page?: int = 1}`	Read content from PDF file for specified page range. Supports both local file paths and URLs. For URLs, the PDF will be downloaded to a temporary directory and cached for future use. Note: Avoid reading too many pages at once (recommended: <50 pages) to prevent errors. Args: pdf_file_path: Path to the PDF file or URL to PDF start_page: Starting page number (default: 1) end_page: Ending page number (default: 1) Returns: Extracted text content from the specified pages
`pdf-tools get-pdf-info`	`{pdf_file_path: str}`	Get basic information about a PDF file including page count. Supports both local file paths and URLs. For URLs, the PDF will be downloaded to a temporary directory and cached for future use. Args: pdf_file_path: Path to the PDF file or URL to PDF Returns: Basic information about the PDF file
`pdf-tools merge-pdfs`	`{pdf_paths: list[str], output_path: str}`	Merge multiple PDF files into one. Supports both local file paths and URLs. URLs will be downloaded to temporary files before merging. Mixed local and URL paths are supported. Args: pdf_paths: List of paths to PDF files to merge (in order) - can include URLs output_path: Path where the merged PDF will be saved (must be local path) Returns: Success message with merge details or error message
`pdf-tools extract-pdf-pages`	`{source_path: str, page_numbers: list[int], output_path: str}`	Extract specific pages from a PDF and create a new PDF. Supports URLs for source PDF. The source PDF will be downloaded to a temporary directory if it's a URL. Output path must be a local file path. Args: source_path: Path to the source PDF file or URL to PDF page_numbers: List of page numbers to extract (1-indexed) output_path: Path where the new PDF will be saved (must be local path) Returns: Success message with extraction details or error message
`pdf-tools search-pdf-content`	`{pdf_file_path: str, pattern: str, page_size?: int = 10}`	Search for regex pattern in PDF content and return paginated results. Supports both local file paths and URLs. For URLs, the PDF will be downloaded to a temporary directory and cached for future use. Args: pdf_file_path: Path to the PDF file or URL to PDF pattern: Regular expression pattern to search for page_size: Number of results per page (10-50, default: 10) Returns: Search results with UUID for pagination, or error message
`pdf-tools search-pdf-next-page`	`{search_id: str}`	Get next page of search results. Args: search_id: Search session ID from previous search Returns: Next page of search results or error message
`pdf-tools search-pdf-prev-page`	`{search_id: str}`	Get previous page of search results. Args: search_id: Search session ID from previous search Returns: Previous page of search results or error message
`pdf-tools search-pdf-go-page`	`{search_id: str, page_number: int}`	Go to specific page of search results. Args: search_id: Search session ID from previous search page_number: Page number to go to (1-indexed) Returns: Specified page of search results or error message
`pdf-tools search-pdf-info`	`{search_id: str}`	Get information about a search session. Args: search_id: Search session ID from previous search Returns: Information about the search session

excel

Use excel <command> '<JSON arguments>'. Run excel <command> --help to get more information.

Command	Arguments	Description
`excel apply-formula`	`{filepath: str, sheet_name: str, cell: str, formula: str}`	Apply Excel formula to cell. Excel formula will write to cell with verification.
`excel validate-formula-syntax`	`{filepath: str, sheet_name: str, cell: str, formula: str}`	Validate Excel formula syntax without applying it.
`excel format-range`	`{filepath: str, sheet_name: str, start_cell: str, end_cell?: str = null, bold?: bool = false, italic?: bool = false, underline?: bool = false, font_size?: int = null, font_color?: str = null, bg_color?: str = null, border_style?: str = null, border_color?: str = null, number_format?: str = null, alignment?: str = null, wrap_text?: bool = false, merge_cells?: bool = false, protection?: dict = null, conditional_format?: dict = null}`	Apply formatting to a range of cells.
`excel read-data-from-excel`	`{filepath: str, sheet_name: str, start_cell?: str = "A1", end_cell?: str = null, preview_only?: bool = false}`	Read data from Excel worksheet with cell metadata including validation rules. Args: filepath: Path to Excel file sheet_name: Name of worksheet start_cell: Starting cell (default A1) end_cell: Ending cell (optional, auto-expands if not provided) preview_only: Whether to return preview only Returns: JSON string containing structured cell data with validation metadata. Each cell includes: address, value, row, column, and validation info (if any).
`excel write-data-to-excel`	`{filepath: str, sheet_name: str, data: list[list], start_cell?: str = "A1"}`	Write data to Excel worksheet. Excel formula will write to cell without any verification. PARAMETERS: filepath: Path to Excel file sheet_name: Name of worksheet to write to data: List of lists containing data to write to the worksheet, sublists are assumed to be rows start_cell: Cell to start writing to, default is "A1"
`excel create-workbook`	`{filepath: str}`	Create new Excel workbook.
`excel create-worksheet`	`{filepath: str, sheet_name: str}`	Create new worksheet in workbook.
`excel create-chart`	`{filepath: str, sheet_name: str, data_range: str, chart_type: str, target_cell: str, title?: str = "", x_axis?: str = "", y_axis?: str = ""}`	Create chart in worksheet.
`excel create-pivot-table`	`{filepath: str, sheet_name: str, data_range: str, rows: list[str], values: list[str], columns?: list[str] = null, agg_func?: str = "mean"}`	Create pivot table in worksheet.
`excel create-table`	`{filepath: str, sheet_name: str, data_range: str, table_name?: str = null, table_style?: str = "TableStyleMedium9"}`	Creates a native Excel table from a specified range of data.
`excel copy-worksheet`	`{filepath: str, source_sheet: str, target_sheet: str}`	Copy worksheet within workbook.
`excel delete-worksheet`	`{filepath: str, sheet_name: str}`	Delete worksheet from workbook.
`excel rename-worksheet`	`{filepath: str, old_name: str, new_name: str}`	Rename worksheet in workbook.
`excel get-workbook-metadata`	`{filepath: str, include_ranges?: bool = false}`	Get metadata about workbook including sheets, ranges, etc.
`excel merge-cells`	`{filepath: str, sheet_name: str, start_cell: str, end_cell: str}`	Merge a range of cells.
`excel unmerge-cells`	`{filepath: str, sheet_name: str, start_cell: str, end_cell: str}`	Unmerge a range of cells.
`excel copy-range`	`{filepath: str, sheet_name: str, source_start: str, source_end: str, target_start: str, target_sheet?: str = null}`	Copy a range of cells to another location.
`excel delete-range`	`{filepath: str, sheet_name: str, start_cell: str, end_cell: str, shift_direction?: str = "up"}`	Delete a range of cells and shift remaining cells.
`excel validate-excel-range`	`{filepath: str, sheet_name: str, start_cell: str, end_cell?: str = null}`	Validate if a range exists and is properly formatted.
`excel get-data-validation-info`	`{filepath: str, sheet_name: str}`	Get all data validation rules in a worksheet. This tool helps identify which cell ranges have validation rules and what types of validation are applied. Args: filepath: Path to Excel file sheet_name: Name of worksheet Returns: JSON string containing all validation rules in the worksheet

memory

Use memory <command> '<JSON arguments>'. Run memory <command> --help to get more information.

Command	Arguments	Description
`memory create-entities`	`{entities: list[{name: str, entityType: str, observations: list[str]}]}`	Create multiple new entities in the knowledge graph — `entities.name`: The name of the entity; `entities.entityType`: The type of the entity; `entities.observations`: An array of observation contents associated with the entity
`memory create-relations`	`{relations: list[{from: str, to: str, relationType: str}]}`	Create multiple new relations between entities in the knowledge graph. Relations should be in active voice — `relations.from`: The name of the entity where the relation starts; `relations.to`: The name of the entity where the relation ends; `relations.relationType`: The type of the relation
`memory add-observations`	`{observations: list[{entityName: str, contents: list[str]}]}`	Add new observations to existing entities in the knowledge graph — `observations.entityName`: The name of the entity to add the observations to; `observations.contents`: An array of observation contents to add
`memory delete-entities`	`{entityNames: list[str]}`	Delete multiple entities and their associated relations from the knowledge graph — `entityNames`: An array of entity names to delete
`memory delete-observations`	`{deletions: list[{entityName: str, observations: list[str]}]}`	Delete specific observations from entities in the knowledge graph — `deletions.entityName`: The name of the entity containing the observations; `deletions.observations`: An array of observations to delete
`memory delete-relations`	`{relations: list[{from: str, to: str, relationType: str}]}`	Delete multiple relations from the knowledge graph — `relations`: An array of relations to delete; `relations.from`: The name of the entity where the relation starts; `relations.to`: The name of the entity where the relation ends; `relations.relationType`: The type of the relation
`memory read-graph`	`{}`	Read the entire knowledge graph
`memory search-nodes`	`{query: str}`	Search for nodes in the knowledge graph based on a query — `query`: The search query to match against entity names, types, and observation content
`memory open-nodes`	`{names: list[str]}`	Open specific nodes in the knowledge graph by their names — `names`: An array of entity names to retrieve

fetch

Use fetch <command> '<JSON arguments>'. Run fetch <command> --help to get more information.

Command	Arguments	Description
`fetch html`	`{url: str, headers?: dict}`	Fetch a website and return the content as HTML — `url`: URL of the website to fetch; `headers`: Optional headers to include in the request
`fetch markdown`	`{url: str, headers?: dict}`	Fetch a website and return the content as Markdown — `url`: URL of the website to fetch; `headers`: Optional headers to include in the request
`fetch txt`	`{url: str, headers?: dict}`	Fetch a website, return the content as plain text (no HTML) — `url`: URL of the website to fetch; `headers`: Optional headers to include in the request
`fetch json`	`{url: str, headers?: dict}`	Fetch a JSON file from a URL — `url`: URL of the JSON to fetch; `headers`: Optional headers to include in the request

web-search

Use web-search <command> '<JSON arguments>'. Run web-search <command> --help to get more information.

Command	Arguments	Description
`web-search web-search`	`{query: str, num_results?: int = 10}`	Search the web using Google Serper API with concurrency control and retry mechanisms. Supports various Google search operators. — `query`: Search query with optional Google search operators.; `num_results`: Number of results to return, default 10, max 50

python3

Execute Python code directly under the agent workspace and return stdout and stderr. Code can be pasted into a .py file and run with python3.

Python cannot launch subprocesses or make direct network requests. Run the documented service commands directly in the shell. Use Python for local file processing and generating JSON arguments.

VFS passed: 4/5
MCP passed: 3/5
Fewer tokens with VFS: 4 of 5 models

Passes and mean tokens per run for each model, VFS vs MCP
Model	VFS	MCP	VFS Tokens	MCP Tokens	Token Reduction Rate
Claude Opus 5.5	1/1	0/1	55.1k	147k	62%
GPT-6.1 Sol	1/1	1/1	136k	197k	31%
GPT-5.6 Luna	1/1	1/1	356k	272k	−31%
Claude Haiku 5.5	1/1	1/1	1.14M	1.58M	28%
Gemini 3.8 Flash	0/1	0/1	2.64M	7.56M	65%