Data Analyst

Enterprise-grade data analysis assistant. Clean, analyze, and visualize data automatically. **Triggers when user mentions:** - Data cleaning: "数据清洗", "整理数据",...

Security Scan

The skill's documentation, dependencies, and install script are plausible for a local data-analysis tool, but there are inconsistencies in the registry metadata and truncated/missing code inspection (test.sh and the tools/*.py sources were not provided), so the runtime behavior cannot be fully verified.

Purpose & Capability

The skill name, description, required binaries (python3, pip3), and listed Python deps (pandas, openpyxl, matplotlib) align with a local data analysis tool. However: (1) registry metadata at the top claims "No install spec — instruction-only skill" while SKILL.md includes an install section and installScripts (install.sh), and code files are present — this is inconsistent; (2) marketing/pricing mention enterprise features (API access, SSO, hosted signup) that are not implemented or justified by the packaged code and docs. These mismatches are worth clarifying.

Instruction Scope

SKILL.md instructs local processing and shows only local commands (calling tools/analyze.py etc.), which is appropriate. But the runtime install script runs test.sh and the repository contains executable Python tools (tools/*.py) whose full sources were omitted from the provided content. Because test.sh and the tools code were not shown, we cannot confirm they confine themselves to local data — test scripts or tools could perform network I/O or other unexpected actions. Documentation includes example commands that optionally upload or email reports (aws s3 cp, mail) — these are examples, not necessarily executed by the skill, but they indicate possible integration points that could be used to send data externally if enabled.

Install Mechanism

The install flow uses pip3 to install common Python packages (pandas, openpyxl, matplotlib, seaborn) and an install.sh script to run tests and chmod files. This is a standard packaging approach and does not pull code from obscure URLs or use archive extraction. Risk is moderate because pip installs packages from PyPI (expected for Python projects), and install.sh executes test.sh (contents not shown), so the script should be reviewed before running.

Credentials

The skill requests no environment variables or credentials. It creates/uses a per-skill config path (~/.openclaw/skills/data-analyst/config.json) which is reasonable. There are no declared credentials or sensitive env vars required by the skill, so requested environment access appears proportionate to the declared purpose.

Persistence & Privilege

The skill is not marked always:true and does not claim or appear to modify other skills or global agent settings. install.sh sets permissions and runs tests inside the skill folder; that is standard for self-installing skills and not privileged beyond the user's account. No autonomous invocation flags are unusual beyond the platform defaults.

What to consider before installing

This skill mostly looks like a normal local Python data-analysis tool, but you should not install or run it blind. Before installing: (1) Ask for and review the full contents of test.sh and all tools/*.py to ensure there are no network calls, telemetry, or commands that exfiltrate files or credentials. (2) Inspect install.sh and test.sh locally (do not run them) or run them in a disposable/sandboxed environment (container or VM). (3) Confirm the apparent mismatch in registry metadata (it says "instruction-only" but includes install scripts and code) and ask the publisher to explain enterprise features (API/SSO) referenced in marketing — if those exist, review how credentials are handled. (4) If you must run it on sensitive data, run a code-only dry run (static review) or run it against synthetic data first, and ensure no automatic uploads (S3, mail, analytics) are triggered. If you can't review the omitted Python files, treat the package as untrusted.

Data Analyst Skill

Automatically clean, analyze, and visualize enterprise data.

Features

Feature	Description	Reference
Data Cleaning	Remove duplicates, handle missing values, standardize formats	references/data_cleaning.md
Data Analysis	Statistics, trends, correlations	references/data_analysis.md
Visualization	Charts, graphs, dashboards	references/visualization.md
Report Generation	Automated insights and recommendations	references/report_generation.md

Quick Start

Step 1: Prepare Your Data

Place your data file (Excel/CSV/JSON) in a known location.

Step 2: Analyze Data

# Basic analysis
{baseDir}/tools/analyze.py data.csv

# With specific options
{baseDir}/tools/analyze.py data.xlsx --clean --visualize --report

Step 3: Get Results

Output includes:

• Cleaned data file
• Analysis summary
• Visualization charts
• Insights report

Available Tools

Tool	Function	Input	Output
analyze.py	Main analysis entry point	Data file	Summary + options
clean.py	Data cleaning	Raw data	Clean data
visualize.py	Generate charts	Data	PNG/PDF charts
report.py	Generate reports	Analysis results	Markdown report

Usage Examples

Example 1: Quick Analysis

"帮我分析这个销售数据"

# Place your file as sales_data.csv
{baseDir}/tools/analyze.py sales_data.csv

Output:

✅ Data loaded: 1,234 rows, 8 columns
📊 Summary statistics generated
📈 Visualization: sales_trend.png
💡 3 key insights found

Example 2: Data Cleaning + Analysis

"清洗并分析客户数据"

{baseDir}/tools/analyze.py customer_data.xlsx --clean --visualize

Example 3: Generate Full Report

"生成完整的数据报告"

{baseDir}/tools/analyze.py data.csv --report --output report.md

Supported Formats

Format	Read	Write	Notes
CSV	✅	✅	Universal format
Excel (.xlsx)	✅	✅	Requires openpyxl
JSON	✅	✅	Structured data
TSV	✅	✅	Tab-separated

Output Files

File	Description
*_cleaned.csv	Cleaned data
*_summary.txt	Statistical summary
*_chart_*.png	Visualizations
*_report.md	Full analysis report

Common Use Cases

Business Analytics

• Sales trend analysis
• Customer segmentation
• Revenue forecasting
• Performance dashboards

Data Quality

• Duplicate detection
• Missing value handling
• Format standardization
• Anomaly detection

Reporting

• Executive summaries
• Department reports
• Trend analysis
• KPI tracking

Advanced Features

Custom Analysis

# Specific columns only
{baseDir}/tools/analyze.py data.csv --columns "sales,date,region"

# Time series analysis
{baseDir}/tools/analyze.py data.csv --timeseries --date-column "date"

# Group by category
{baseDir}/tools/analyze.py data.csv --group-by "region" --aggregate "sum,mean"

Visualization Options

# Chart types
{baseDir}/tools/visualize.py data.csv --type bar
{baseDir}/tools/visualize.py data.csv --type line
{baseDir}/tools/visualize.py data.csv --type scatter
{baseDir}/tools/visualize.py data.csv --type pie

# Styling
{baseDir}/tools/visualize.py data.csv --style professional
{baseDir}/tools/visualize.py data.csv --colors "blue,green,red"

Setup

# Install dependencies
pip3 install pandas openpyxl matplotlib seaborn

# Verify installation
python3 -c "import pandas, matplotlib; print('Dependencies OK')"

Notes

• ⚠️ Large files (>100MB) may take time to process
• ⚠️ Excel files require openpyxl
• ⚠️ Charts saved as PNG by default
• ⚠️ All processing is local (no data sent externally)

Troubleshooting

"Module not found"

pip3 install pandas openpyxl matplotlib

"File encoding error"

• Try converting to UTF-8 first
• Or specify encoding: --encoding gbk

"Memory error with large files"

• Process in chunks: --chunk-size 10000
• Or sample data: --sample 0.1