Log messages to both console and file

Log messages to both console and file

Print message to terminal log file (and optionally console) This captures the exact formatted output with emojis Arg

Log message to terminal

Log messages to both console and file

Log error message Args: message: Error message context: Additional context dictionary ex

Track API call cost (e.g., JINA search, OCR) based on token usage Args: tokens: Number of tokens used price_

End tracking for current task and save consolidated task record

Start tracking costs for a new task Args: task_id: Unique identifier for the task date: Date task was assign

Track token usage and calculate cost Args: input_tokens: Number of input tokens output_tokens: Number of out

Add income from completed work with evaluation score threshold Payment is only awarded if evaluation_score >= min_evaluation

Log warning message Args: message: Warning message context: Additional context dictionary

Log messages with timestamp

Log info message Args: message: Info message context: Additional context dictionary prin

Initialize tracker, load existing state or create new

Get current balance

Run the wrap-up workflow Args: date: Current date (YYYY-MM-DD) task: Current task dictionary (may be

Convenience function to log error

Get survival status based on balance Returns: Status: "thriving", "stable", "struggling", or "bankrupt"

* Build API Gateway-compatible response

Get or create the singleton session sandbox instance.

Get cost breakdown for a specific task Args: task_id: Task identifier Returns:

Get task by ID from filtered task list Args: task_id: Task identifier Returns: Task dictionary or N

Get the global logger instance

Load tasks from configured source (parquet, jsonl, or inline) Returns: Number of tasks loaded Raises:

Select a task for the given date Args: date: Date string (YYYY-MM-DD) signature: Agent signature (optional,

Write log entry to file

Close all server connections

Get cost summary for a specific date Args: date: Date string (YYYY-MM-DD) Returns:

Convert category name to safe filename

Check if agent is bankrupt

Apply agent-specific filters to task list

Log evaluation result Args: artifact_path: Single path (str) or list of paths

Broadcast message to all connected clients

Log debug message Args: message: Debug message context: Additional context dictionary pr

Evaluate a work artifact and determine payment Args: signature: Agent signature task: Task dictionary (must

Extract previewable artifacts from a task's evaluation data

Get detailed cost analytics across all tasks and dates Returns: Dictionary with cost breakdown by channel, date,

Get total daily token cost

Get net worth (balance + portfolio value)

Ensure a provider backend is ready and return its native handle.

Get comprehensive economic summary Returns: Dictionary with all economic metrics

Load configuration from JSON file

Load all records from a JSONL file; skip malformed lines.

Load a JSONL file, return list of dicts.

Convert PDF to list of PNG images, combining 4 pages into one image to save resources. Uses pdf2image to convert PDF pages to images, th

Convert PPTX to list of PNG images (one per slide). Uses LibreOffice to convert PPTX → PDF, then pdf2image to convert PDF → PNG images.

Save end-of-day economic state Args: date: Date string (YYYY-MM-DD) work_income: Today's work income (actual

Exit with an error if clawwork is not enabled in config.

Return a safe default classification.

Return a short cost footer for the current task.

Get global state from parent module

Load task_completions.jsonl indexed by task_id → entry dict.

Log messages to log file

Log task assignment to agent's task log

Create a ClawWorkAgentLoop from nanobot config. Shared by both the ``agent`` and ``gateway`` commands. Returns ``(agent_loop, state, bus)``.

Print final simulation summary

Extract text content from DOCX file Args: path: Path to DOCX file Returns: Extr

Extract data from XLSX file Args: path: Path to XLSX file Returns: Formatted te

Save balance record to file

Validate task has required fields Args: task: Task dictionary identifier: Line number or index for error mes

Clean up session sandbox resources.

Use GPT-4o to generate evaluation guidelines for a specific category