onesie

Overview

onesie is a sophisticated and expressive command-line interface (CLI) designed for Unix systems, specifically tailored for working with System One models, such as TypeSafe's Jev and Berget's models. This tool facilitates seamless interaction with natural language processing capabilities, allowing users to pipe text into the command line and receive actionable responses—ideal for automating decision-making processes and scripting.

Key Features

Text Processing

Effortlessly pipe text into onesie to query its capabilities and receive an assessment based on urgency or relevance. For instance:

echo 'EVERYTHING IS DOWN, CALL ME NOW' | onesie 'does this convey urgency' -r
# Response: 0.99

Command Gating

Manage shell commands safely with integrated exit codes that determine actions:

onesie -f shell-safety -q --state 'git status'
# exit 0: run it
onesie -f shell-safety -q --state 'rm -rf ~'
# exit 1: block it
onesie -f shell-safety -q --state 'echo hello > notes.txt'
# exit 7: not sure, ask a person

Evidence Calibration

With the calibrate feature, assess accuracy by analyzing prior labeled records, enabling fine-tuning of the system's responses. An example command would be:

onesie calibrate -f shell-safety -i jsonl --map .command --id .id -m jev-1.13.0 \
    --label destroys=.destroys --label secrets=.secrets --label network=.network \
    --cuts 0.25,0.5,0.65 --out answers.jsonl --resume < commands.jsonl

Intelligent Questioning

The CLI allows for complex question formations and provides rate-based assessments of inquiries:

echo 'I was charged twice for my order' | onesie 'who should handle this' --pick billing,shipping,technical -r
# Response: billing

Workflow Integration

Integrate onesie into scripts to enhance automated workflows and decision-making processes, for instance, in a triage ticketing system:

tail -f tickets.jsonl | onesie 'is this urgent' -i jsonl --map .body --id .id -o json

This functionality can also extend to handling pull requests on GitHub, ensuring vital communications regarding code changes are automated and streamlined.

Use Cases

• Triage Tickets: Automate the assessment of incoming tickets, determining their urgency in real time.
• Gate Scripts: Implement decision gates that restrict command execution based on confidence scoring.
• Batch Processing: Process data in bulk while keeping track of completed and pending inquiries efficiently.

Additional Information

onesie supports various providers including TypeSafe, OpenRouter, and Berget, providing flexibility in how queries are routed. It also includes features like caching for repeated requests, dry runs for testing requests without execution, and a comprehensive help system for user guidance.

For detailed command references, see REFERENCE.md. For contributors, guidelines are available in CONTRIBUTING.md.