Is this compatible with legacy tools?

No, this skill is specifically designed to use modern hive and swarm tools while explicitly avoiding deprecated bd and cass references.

What is the Always-On Guidance skill?

It is a framework that provides persistent rules and model-specific defaults for agents operating within a swarm environment to ensure consistent behavior.

How does it handle different AI models?

It differentiates between models like GPT-5.2-code (prioritizing strict checklists) and Opus 4.5 (allowing brief rationales) to maximize the efficiency of each model type.

Does this skill support automated testing?

Yes, it enforces a 'Red-Green-Refactor' discipline whenever tests cover the area being modified by a swarm agent.

What is the purpose of file reservation?

File reservation allows agents to 'lock' specific files with a TTL (Time To Live) to prevent multiple agents in a swarm from making conflicting changes simultaneously.

Always-On Swarm Guidance

Name: Always-On Swarm Guidance
Author: joelhooks

byjoelhooks

•

359

협업 도구

Enforces strict behavioral rules and model-specific logic for multi-agent swarm coordination and tool usage.

소개

This skill provides a persistent framework for multi-agent systems using the Swarm Plugin, ensuring all sub-agents adhere to specific instruction hierarchies and tool usage patterns. It optimizes performance by tailoring output styles for specific models—such as GPT-5.2-code and Opus 4.5—while maintaining strict file-handling protocols and testing disciplines to prevent regressions. By centralizing guidance, it ensures that swarm agents remain focused, concise, and aligned with project-specific development standards during complex, collaborative coding tasks.

주요 기능

Automated swarm subtask spawning and task handoff management
Instruction priority enforcement across system, developer, and agent levels
Integrated Red-Green-Refactor testing discipline enforcement
Model-specific behavior tuning for GPT-5.2-code and Opus 4.5
359 GitHub stars
File reservation system with TTL-based locking to prevent concurrent edit collisions

사용 사례

Managing concurrent file edits in complex repositories to ensure state consistency
Standardizing AI response styles and reasoning depths across different LLM backends
Coordinating large-scale refactors across multiple files using a swarm of autonomous agents

소개

주요 기능

Automated swarm subtask spawning and task handoff management
Instruction priority enforcement across system, developer, and agent levels
Integrated Red-Green-Refactor testing discipline enforcement
Model-specific behavior tuning for GPT-5.2-code and Opus 4.5
359 GitHub stars
File reservation system with TTL-based locking to prevent concurrent edit collisions

사용 사례

Managing concurrent file edits in complex repositories to ensure state consistency
Standardizing AI response styles and reasoning depths across different LLM backends
Coordinating large-scale refactors across multiple files using a swarm of autonomous agents