GitPatch Goals CLI

SkillDev tools

Lets your agent check whether generated code actually meets the original goals and coding standards.

Available today. Use it from your connected AI after setup.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the GitPatch Goals CLI skill

About this skill

Evaluates generated code via GitPatch against original goals and coding standards.

What this skill tells your AI

The instructions your AI receives, as published by google-labs-code/jules-sdk in packages/core/examples/gitpatch-goals/SKILL.md and read by ahel’s review.

This CLI uses Jules sessions to simulate code generation, extracts the resulting code as a GitPatch, and feeds it to a second Jules session to evaluate if it successfully met the original prompt's goals and adheres to coding standards.

Usage Guidelines for AI Agents

When invoking this CLI, adhere to the following best practices:

  1. Schema Introspection: You can introspect the required arguments and schema at runtime by passing --describe.

    bun run index.ts --describe
    
  2. Context Window Discipline: Use --json for predictable, deterministic, machine-readable output. Avoid parsing raw terminal stdout.

    bun run index.ts --prompt "Create an API" --json
    
  3. Input Hardening: Before executing mutations or relying on long-running APIs (like creating Jules Sessions), validate your payload using the --dry-run flag to ensure the CLI safely accepts your arguments without executing side effects.

    bun run index.ts --prompt "Create an API" --dry-run
    

Signals

GitHub stars
134
Forks
32
Last commit
Jul 2026

ahel review

  • K6info
    bundled executables the agent is told to run

Automated review, not a security audit. Ruleset v1+k2.

Advanced
Catalog kind
skill
Key
gitpatch-goals-cli
Source
github.com/google-labs-code/jules-sdk