name: filter-csv-columns description: Filter a CSV file to keep only columns relevant for recruiter screening. Consolidates jobs, education, and skills into clean columns. argument-hint: [input-csv-path] [output-csv-path] allowed-tools: Read, Bash(python:*), Write

Filter CSV Columns for Screening

Filter an enriched LinkedIn CSV to keep only columns relevant for recruiter screening.

Task

Input file: $ARGUMENTS[0] (or ask user if not provided) Output file: $ARGUMENTS[1] (or default to filtered_profiles.csv in the same folder)

Output Columns (14 total)

1. first_name - First name
2. last_name - Last name
3. headline - LinkedIn headline
4. location - Location
5. current_title - Current job title (from job_1_job_title)
6. current_company - Current company (from job_1_job_company_name)
7. current_start_date - When current role started (from job_1_job_start_date)
8. current_years_in_role - Years in current role (CALCULATED from current_start_date to today)
9. current_description - Current job description (from job_1_job_description)
10. summary - Profile summary/bio
11. past_positions - All past jobs combined: "Title at Company (start - end) [X yrs]: description || ..."
12. education - All education combined: "School, Degree in Field | ..."
13. skills - All skills combined: "Skill1, Skill2, Skill3, ..."
14. public_url - LinkedIn profile URL

Logic

Current Position

• current_title = job_1_job_title
• current_company = job_1_job_company_name
• current_start_date = job_1_job_start_date
• current_years_in_role = CALCULATE from current_start_date to today (don't use job_1_years_in_job as it may be outdated)
• current_description = job_1_job_description

Past Positions (job_2, job_3, etc.)

Combine into single past_positions column with format:

Title at Company (start_date - end_date) [X yrs]: description || Title at Company (dates) [X yrs]: description

• Use || as separator between jobs
• CALCULATE years from start_date to end_date for each position
• Include description if available
• Skip job_1 (that's current position)

Education (edu_1, edu_2, etc.)

Combine into single education column with format:

School Name, Degree in Field of Study | School Name, Degree in Field

• Use | as separator between schools

Skills (skill_1_name, skill_2_name, etc.)

Combine ALL skill_N_name columns into single skills column:

Skill1, Skill2, Skill3, ...

• Comma-separated
• Include all skills (not just first 10)

Example Usage

/filter-csv-columns "C:\data\raw_profiles.csv" "C:\data\filtered_profiles.csv"

Python Implementation

import pandas as pd
import re
from datetime import datetime

df = pd.read_csv(input_file)

# Find all job/edu/skill column numbers
job_nums = sorted(set(int(re.match(r'^job_(\d+)_', col).group(1)) for col in df.columns if re.match(r'^job_(\d+)_', col)))
edu_nums = sorted(set(int(re.match(r'^edu_(\d+)_', col).group(1)) for col in df.columns if re.match(r'^edu_(\d+)_', col)))
skill_cols = [col for col in df.columns if re.match(r'^skill_\d+_name$', col)]

# Function to calculate years from date string to today
def calc_years_to_today(date_str):
    if pd.isna(date_str) or not str(date_str).strip():
        return None
    try:
        dt = datetime.strptime(str(date_str).strip(), '%d %b %Y')
        years = (datetime.now() - dt).days / 365.25
        return round(years, 1)
    except:
        return None

# Current position - CALCULATE years from start date
df['current_title'] = df.get('job_1_job_title', '')
df['current_company'] = df.get('job_1_job_company_name', '')
df['current_start_date'] = df.get('job_1_job_start_date', '')
df['current_years_in_role'] = df['job_1_job_start_date'].apply(calc_years_to_today)
df['current_description'] = df.get('job_1_job_description', '')

# Combine functions for skills, jobs, education
# - For past positions: calculate years from start_date to end_date
# - For skills: combine all skill_N_name columns
# - For education: combine school, degree, field

# Final columns
cols = ['first_name', 'last_name', 'headline', 'location', 'current_title',
        'current_company', 'current_start_date', 'current_years_in_role',
        'current_description', 'summary', 'past_positions', 'education',
        'skills', 'public_url']
df_filtered = df[cols]
df_filtered.to_csv(output_file, index=False)