Skip to content

ajstafford/jules-ai-cli-skill

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 

Repository files navigation

Jules CLI Skill

Overview

This skill enables agents to interact with the Jules CLI for task assignment, session monitoring, and pulling patches from completed remote sessions. It is designed with a Security-First & Manual philosophy to prevent runaway automation and ensure resources are used only for complex, appropriate tasks.

Key Features:

  • Security-First Design - Mandatory input validation and shell injection prevention (via quoting and environment variables).
  • Mandatory Approvals - Explicit user confirmation required for session creation and code modifications.
  • Safety Controls - Built-in complexity thresholds and anti-proliferation rules.
  • Direct CLI Control - Leverages the raw jules CLI for maximum transparency.
  • Robust Parsing - Safe Python one-liners for status parsing without external scripts.

Installation

  1. Clone or download this repository.
  2. Install Jules CLI and authenticate:
    npm install -g @google/jules
    jules login
  3. Verify setup:
    jules remote list --repo
  4. Point your agent to the jules-cli directory. The agent will read SKILL.md to understand its security and operating boundaries.

Usage Guidelines (CRITICAL)

To prevent inappropriate resource usage, the following rules are enforced:

  • Local First: Use local tools for small refactors, comments, or typos.
  • Complexity Threshold: Only use Jules for large-scale, isolated, or exploratory tasks.
  • No Proliferation: One session at a time. Never bulk-create or loop over session creation.
  • Mandatory Approval: The agent must ask for permission before creating remote VMs or applying code changes.

Documentation

  • SKILL.md - Main skill definition with security rules, mandatory safety controls, and metadata.
  • usage.md - Secure command reference for native Jules CLI commands.

Project Structure

jules-cli/
├── SKILL.md              # Security guidelines, safety controls, and metadata
├── references/
│   └── usage.md          # Secure command usage reference

Security & Prerequisites

  1. Jules CLI - Must be installed (jules) and authenticated.
  2. Python 3 - Required for safe status parsing utilities.
  3. Environment - The skill requires HOME access for credential management.
  4. GitHub Account - Required for Jules remote sessions.

License

This project is provided for educational and demonstration purposes. Feel free to use, modify, and distribute with attribution.

About

A specialized Agent Skill that enables local AI agents to orchestrate remote Jules cloud tasks

Topics

Resources

Stars

Watchers

Forks

Contributors