always-on-guidance

Enforce rule-based guidance for Claude plugin agents across sessions.

729|67|Updated Dec 8, 2025

Search Tags:#guidance #swarm #agent-coordination #claude-plugin #always-on #rule-based

Ready to run

npx skills add https://github.com/joelhooks/swarm-tools --skill always-on-guidance

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: always-on-guidance
Source: https://github.com/joelhooks/swarm-tools/tree/main/packages/claude-code-swarm-plugin/skills/always-on-guidance
Command: npx skills add https://github.com/joelhooks/swarm-tools --skill always-on-guidance

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a stable, always-on guidance layer for claude-plugin agents to align behavior, tool usage, and model-defaults across tasks.

Core Features & Use Cases

Rule-based guidance: Enforces consistent decision-making and tool usage for Claude Code/OpenCode sessions.
Swarm integration: Coordinates with swarm-plugin tools for governance and progress tracking.
Risk reduction: Reduces drift by applying standardized defaults and prohibiting deprecated references.

Quick Start

Activate always-on guidance during a Claude Code session to enforce the rule set for agents.

Frequently Asked Questions about always-on-guidance

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

FAQPage Schema

How do I enforce consistent rules for Claude Code agents across sessions?▼

To enforce consistent behavior, apply rule-based guidance that codifies defaults and tool usage for Claude Code agents. This provides an always-on layer to align decisions and reduce drift across sessions.

What is always-on guidance for multi-agent swarm coordination?▼

Always-on guidance is a stable rule layer that aligns behavior and tool usage for claude-plugin agents. It integrates with swarm plugin tools to govern decisions and coordinate progress tracking across multiple agents.

Can I use this rule-based guidance with OpenCode sessions?▼

Yes, this guidance applies across both Claude Code and OpenCode sessions. It enforces standardized defaults and prohibits deprecated references to maintain consistent agent behavior throughout your workflows.

How do I track multi-agent progress when coordinating swarm plugins?▼

You can track multi-agent progress by relying on dedicated commands for progress logging integrated with the swarm plugin tools. This ensures coordinated governance and consistent task tracking across agents.

Why does my Claude plugin agent behavior drift between different tasks?▼

Agent behavior drifts when standardized defaults and rules are not enforced. Applying an always-on rule-based guidance layer codifies decision-making parameters and tool usage to reduce this drift across tasks.

What are the limitations of using frontmatter to codify agent rules?▼

The limitation is that this basic implementation relies on codifying rules within frontmatter and content sections, meaning complex dynamic logic may not be supported. It provides static, rule-based guidance rather than adaptive behaviors.