jin-fsd

Execute multi-agent AI workflows with step-by-step user approval gates.

2|Updated Mar 3, 2026

Search Tags:#orchestration #ai workflow #user approval #fsd #self-driving #safe execution

Ready to run

npx skills add https://github.com/cjinzy/jin-claude --skill jin-fsd

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: jin-fsd
Source: https://github.com/cjinzy/jin-claude/tree/main/plugins/jin-claude/skills/jin-fsd
Command: npx skills add https://github.com/cjinzy/jin-claude --skill jin-fsd

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides a safe, user-controlled execution environment for complex AI-driven tasks, ensuring every step is reviewed and approved before proceeding.

Core Features & Use Cases

Full Self-Driving Mode: Executes multi-agent pipelines with mandatory user approval at each stage.
Interactive Control: Allows users to approve, reject, or modify tasks at any point.
Use Case: Safely refactor a critical part of your codebase by having the AI propose changes, then reviewing and approving each modification before it's applied.

Quick Start

Initiate the Full Self-Driving mode for the task 'implement the new user authentication flow'.

Frequently Asked Questions about jin-fsd

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I execute multi-agent AI workflows safely with step-by-step approval?▼

You can execute multi-agent AI workflows safely by using a self-driving execution environment that requires explicit user consent at each state transition. This ensures granular control, allowing you to review, approve, or modify proposed changes before they are applied.

What is the best way to control AI code refactoring tasks before they are applied?▼

The best way to control AI code refactoring is to use an interactive workflow that proposes changes and pauses for mandatory user approval. You can reject or modify the AI's proposed modifications at any point, ensuring safe execution for critical codebase updates.

Can I modify or reject tasks proposed by an AI agent during a complex development workflow?▼

Yes, you can modify or reject tasks proposed by an AI agent during complex development workflows. The interactive control mechanism requires explicit user approval before proceeding, allowing you to alter or cancel proposed modifications at any stage of the pipeline.

Does the self-driving AI execution mode work with task planners and interview agents?▼

Yes, the self-driving AI execution mode integrates directly with task planners, interview agents, and SWE agents. This facilitates comprehensive task execution and verification by orchestrating these components within a single user-approved pipeline.

When do I need user approval gates for AI system analysis and execution?▼

You need user approval gates for AI system analysis and execution when running complex, multi-agent pipelines that modify critical systems. These gates ensure safety by pausing the workflow for explicit consent before each state transition or code change is applied.