test-processing-guide

zhenxiay's avatarfrom zhenxiay

Comprehensive guide for AI agent when it is requested to perform following tasks: (1) Summarize text, (2) Sentiment analysis, (3) Translate test in full length.

0stars🔀0forks📁View on GitHub🕐Updated Jan 9, 2026

When & Why to Use This Skill

This Claude skill is a comprehensive text processing suite designed to automate three critical linguistic tasks: text summarization, sentiment analysis, and full-length translation. By leveraging structured workflows and batch processing capabilities, it enables AI agents to efficiently handle large-scale datasets and data tables, transforming raw text into actionable insights and localized content with high precision.

Use Cases

  • Large-Scale Customer Feedback Analysis: Automatically process thousands of customer reviews to extract sentiment scores and summarize key themes for product improvement.
  • Global Documentation Localization: Translate extensive technical manuals or database entries in full length, ensuring consistent communication across different languages.
  • Automated Content Auditing: Scan large volumes of corporate documents to generate executive summaries and assess the overall tone or sentiment of internal communications.
  • Data Table Enrichment: Process specific text columns within DuckDB or other data tables to add summarized versions or translated equivalents for better data accessibility.
nametest-processing-guide
description"Comprehensive guide for AI agent when it is requested to perform following tasks: (1) Summarize text, (2) Sentiment analysis, (3) Translate test in full length."

Text Summary, Translation and Sentiment analysis

Overview

A user may ask you to summarize text from a file, run sentiment analysis of text or translate text in full length. You have different tools and workflows available for different tasks.

Workflow Decision Tree

Summarize text

Use "Text summary" workflow

Sentiment analysis

Use "Sentiment analysis" workflow

Translate text in full length

Use "Translate text" workflow

Workflow list

Text summary

  1. Ask user for necessary inputs:

    • input_table_name
    • batch_size
    • text_column
    • length
    • output_table

    If user needs help for some of the input parameters, run following command to get help:

    uv run src/workflow/run_workflow.py --help
    
  2. Run following command to finish the requested task:

    uv run src/workflow/run_workflow.py
    

Sentiment analysis

  1. Ask user for necessary inputs:

    • input_table_name
    • batch_size
    • text_column
    • output_table

    If user needs help for some of the input parameters, run following command to get help:

    uv run src/workflow/run_sentiment_analysis.py --help
    
  2. Run following command to finish the requested task:

    uv run src/workflow/run_sentiment_analysis.py
    

Translate text

  1. Ask user for necessary inputs:

    • input_table_name
    • batch_size
    • text_column
    • output_path

    If user needs help for some of the input parameters, run following command to get help:

    uv run src/workflow/run_full_translation.py --help
    
  2. Run following command to finish the requested task:

    uv run src/workflow/run_full_translation.py
    

Error handling

Script cannot be found: If the Agent gets this error, use powershellcommands to search for the script on the local machine.

Parameters missing or incorrect: If the Agent gets error regarding any missing or incorrect input parameters, repeat step 1 and confirm with user.

IOException: In case of IOException because DuckDB is open in another process, get the PID and stop it in powershell.