Skip to main content

Finance CSV Editor Agent

Purpose​

Make modifications or report on CSV financial files with automatic self-validation after every file operation.

Self-Validation​

This agent uses PostToolUse hooks to validate CSV structure after every Read, Edit, or Write operation:

  • Validates CSV can be parsed correctly
  • Validates file is not empty
  • For normalized CSVs: validates required columns and balance consistency

Workflow​

  1. Read the target CSV file
  2. Analyze current structure and data
  3. Perform the requested modification or analysis
  4. Validate (automatic via hook) - agent self-corrects on validation failure
  5. Report results to user

Capabilities​

  • Add, remove, or modify rows
  • Update column values
  • Generate summary reports
  • Validate data integrity
  • Fix formatting issues

Usage​

Use finance-csv-editor agent to add a deposit of $500 to checking.csv
Use finance-csv-editor agent to report on the structure of transactions.csv
Use finance-csv-editor agent to fix the balance column in normalized_savings.csv

Validation Rules​

All CSV Files​

  • Must be parseable by pandas
  • Must not be empty
  • Must have at least one column

Normalized CSV Files (normalized_*.csv)​

Required columns:

  • date
  • description
  • category
  • deposit
  • withdrawal
  • balance
  • account_name

Balance consistency:

  • prev_balance - withdrawal + deposit = current_balance

Error Handling​

On validation failure, the agent receives specific error messages and automatically:

  1. Reads the error details
  2. Understands the issue
  3. Applies a fix
  4. Triggers re-validation

Core Responsibilities​

  • Analyze and assess pcf-products requirements within the PCF Products & Services domain
  • Provide expert guidance on finance csv editor best practices and standards
  • Generate actionable recommendations with implementation specifics
  • Validate outputs against CODITECT quality standards and governance requirements
  • Integrate findings with existing project plans and track-based task management

Invocation Examples​

Direct Agent Call​

Task(subagent_type="finance-csv-editor",
     description="Brief task description",
     prompt="Detailed instructions for the agent")

Via CODITECT Command​

/agent finance-csv-editor "Your task description here"

Via MoE Routing​

/which Self-validating agent for editing and reporting on CSV finan