Datasets

Browse and load ready-to-use AI/ML datasets with fast manipulation. Use when searching datasets, loading training data, transforming formats.

Security Scan

The skill's code and instructions are consistent with a local-only dataset logging/utility: it reads/writes logs under ~/.local/share/datasets, requires no credentials or external installs, and its behavior matches the description.

Purpose & Capability

Name/description (dataset ingestion, transforms, query, export) align with what is provided: a bash CLI that logs operations into per-command .log files under ~/.local/share/datasets. There are no extra binaries, cloud credentials, or unrelated capabilities requested.

Instruction Scope

SKILL.md instructs the agent to run local CLI commands and explains local storage and exports. The included script only reads/writes files inside the declared data directory and uses common Unix utilities; it does not read unrelated system files, access network endpoints, or attempt to exfiltrate environment variables.

Install Mechanism

There is no install spec (instruction-only skill) and the provided code is a single bash script. Nothing is downloaded at install time and no archives or remote installers are used.

Credentials

The skill declares no required env vars, credentials, or config paths. The script uses HOME to determine a local data directory (expected) and does not request or use secrets or external service keys.

Persistence & Privilege

The skill is not always-enabled and is user-invocable. It does not modify other skills or global agent settings. Its persistence is limited to creating and updating files under the user's ~/.local/share/datasets directory.

Assessment

This tool is local-only and appears safe to use, but review the script before installing: it will create and append logs under ~/.local/share/datasets (history.log and per-command .log files) so don't log sensitive secrets into it. The JSON export code does not escape values and may produce invalid JSON if entries contain quotes/newlines — treat exports as potentially containing raw user data. Also confirm you have the full, untruncated script (the provided snippet looked cut off mid-function in the listing); if you plan to allow autonomous invocation, be aware it can write files to your home directory whenever invoked.

A data processing toolkit for ingesting, transforming, querying, and managing dataset entries from the command line. All operations are logged with timestamps and stored locally.

Commands

Data Operations

Each data command works in two modes: run without arguments to view recent entries, or pass input to record a new entry.

Command	Description
datasets ingest <input>	Ingest data — record a new ingest entry or view recent ones
datasets transform <input>	Transform data — record a transformation or view recent ones
datasets query <input>	Query data — record a query or view recent ones
datasets filter <input>	Filter data — record a filter operation or view recent ones
datasets aggregate <input>	Aggregate data — record an aggregation or view recent ones
datasets visualize <input>	Visualize data — record a visualization or view recent ones
datasets export <input>	Export data — record an export entry or view recent ones
datasets sample <input>	Sample data — record a sample or view recent ones
datasets schema <input>	Schema management — record a schema entry or view recent ones
datasets validate <input>	Validate data — record a validation or view recent ones
datasets pipeline <input>	Pipeline management — record a pipeline step or view recent ones
datasets profile <input>	Profile data — record a profile or view recent ones

Utility Commands

Command	Description
datasets stats	Show summary statistics — entry counts per category, total entries, disk usage
datasets export <fmt>	Export all data to a file (formats: json, csv, txt)
datasets search <term>	Search all log files for a term (case-insensitive)
datasets recent	Show last 20 entries from activity history
datasets status	Health check — version, data directory, entry count, disk usage, last activity
datasets help	Show available commands
datasets version	Show version (v2.0.0)

Data Storage

All data is stored locally at ~/.local/share/datasets/:

• Each data command writes to its own log file (e.g., ingest.log, transform.log)
• Entries are stored as timestamp|value pairs (pipe-delimited)
• All actions are tracked in history.log with timestamps
• Export generates files in the data directory (export.json, export.csv, or export.txt)

Requirements

• Bash (with set -euo pipefail)
• Standard Unix utilities: date, wc, du, grep, tail, cat, sed
• No external dependencies or API keys required

When to Use

• To log and track data processing operations (ingest, transform, query, etc.)
• To maintain a searchable history of data pipeline activities
• To export accumulated records in JSON, CSV, or plain text format
• As part of larger automation or data-pipeline workflows
• When you need a lightweight, local-only dataset operation tracker

Examples

# Record a new ingest entry
datasets ingest "loaded training_data.csv 10000 rows"

# View recent transform entries
datasets transform

# Record a query
datasets query "filter by date > 2026-01-01"

# Search across all logs
datasets search "training"

# Export everything as JSON
datasets export json

# Check overall statistics
datasets stats

# View recent activity
datasets recent

# Health check
datasets status

---

Powered by BytesAgain | bytesagain.com | [e-mail] 💬 Feedback & Feature Requests: https://bytesagain.com/feedback