scan

name	scan
description	Quick scan & profile of a content library dataset — ingests, cleans, and generates a data inventory report
argument-hint	<dataset-folder-name>
risk	safe
user-invocable	true
allowed-tools	["Read","Write","Edit","Bash","Glob","Grep"]
model	claude-sonnet-4-6
context	fork
agent	general-purpose

Scan — Content Library Quick Profile

You are executing the :scan skill for the 10x-content-intel plugin.

What This Skill Does

Performs a quick scan and profile of a content library dataset. This is the foundational step — understand the data before analyzing it.

Instructions

Step 1: Locate the Data

• Look in input/{argument}/ for data files (CSV, Excel, JSON)
• If no argument provided, scan all subdirectories in input/
• Read the agent instructions from agents/content-profiler.md

Step 2: Ingest & Profile

Run the following Python script:

import pandas as pd
import os
from pathlib import Path

# Find the dataset
dataset_name = "{argument}"
input_dir = Path(f"input/{dataset_name}")
output_dir = Path(f"output/{dataset_name}")
output_dir.mkdir(parents=True, exist_ok=True)

# Load all data files
files = list(input_dir.glob("*.csv")) + list(input_dir.glob("*.xlsx")) + list(input_dir.glob("*.json"))

for f in files:
    if f.suffix == '.csv':
        df = pd.read_csv(f)
    elif f.suffix in ['.xlsx', '.xls']:
        df = pd.read_excel(f)
    elif f.suffix == '.json':
        df = pd.read_json(f)

    print(f"\n{'='*60}")
    print(f"FILE: {f.name}")
    print(f"{'='*60}")
    print(f"Shape: {df.shape[0]:,} rows x {df.shape[1]} columns")
    print(f"\nColumn Types:\n{df.dtypes.to_string()}")
    print(f"\nMissing Values:\n{df.isnull().sum().to_string()}")
    print(f"\nSample (first 3 rows):")
    print(df.head(3).to_string())
    print(f"\nUnique counts per column:")
    for col in df.columns:
        print(f"  {col}: {df[col].nunique()} unique")

Step 3: Clean the Data

Follow the Content Profiler agent instructions:

• Handle missing values (fill categorical with "Unknown", dates as NaT, numerical with median)
• Remove duplicate rows
• Standardize column names (lowercase, underscores)
• Parse date columns
• Strip whitespace from string columns
• Save cleaned data to output/{argument}/{argument}_cleaned.csv

Step 4: Generate Scan Report

Create output/{argument}/scan_report.md with:

• Dataset overview (source, rows, columns)
• Column inventory table
• Missing data summary
• Data quality score
• Cleaning actions taken
• Key observations

Output

• output/{argument}/scan_report.md
• output/{argument}/{argument}_cleaned.csv