Plugin Settings

Store per-project Claude plugin configuration in YAML frontmatter markdown files.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/stellarone/evaluate --skill plugin-settings-stellarone
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Plugin Settings
Source: https://github.com/stellarone/evaluate/tree/main/.agents/skills/plugin-settings
Command: npx skills add https://github.com/stellarone/evaluate --skill plugin-settings-stellarone

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

Plugin Settings solves the problem of making AI-assisted developer plugins controllable and stateful without editing plugin code, by using a user-local, per-project settings file.

Core Features & Use Cases

  • Per-project plugin configuration & state stored in .claude/plugin-name.local.md so behavior can vary by repository.
  • YAML frontmatter-driven settings (e.g., enabled, validation modes, limits) that hooks, commands, and agents can read deterministically.
  • Operational patterns for safe parsing and defaults, including frontmatter extraction, enabled/disabled gating, and documenting the restart requirement.

What is it for?

Use this skill when you need to store user-configurable plugin behavior such as enabling/disabling a plugin, setting strict vs standard validation, specifying retry limits, or coordinating multi-agent workflows through shared state in the project.

Quick Start

Ask: "Create a .claude/my-plugin.local.md file for this project with enabled: true and a strict_mode: false setting, and explain how a hook should read the YAML frontmatter to decide whether to run."

Frequently Asked Questions about Plugin Settings

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

FAQPage Schema
How do I configure per-repo Claude plugin settings without editing plugin code?

You can configure per-repo Claude plugin settings by creating a `.claude/plugin-name.local.md` file with YAML frontmatter. This pattern stores user-local configuration and runtime state so behavior varies by repository without code changes.

How does YAML frontmatter parsing work for enabling and disabling Claude hooks?

YAML frontmatter parsing for Claude hooks works by extracting settings like `enabled` or `strict_mode` from the markdown file header. Hooks read these flags deterministically to decide whether to run, gating execution based on the parsed configuration values.

Can I coordinate multi-agent workflows using a shared project state file?

Yes, you can coordinate multi-agent workflows by storing shared state in the per-project `.claude/plugin-name.local.md` file. Agents read the YAML frontmatter to synchronize operational behavior and manage workflow coordination across the repository.

What is the best way to customize validation defaults for Claude commands per repository?

The best way to customize validation defaults is by defining settings like `strict_mode` and retry limits in the YAML frontmatter of your per-project local markdown file. Commands read these values to apply repository-specific validation behaviors.

Do I need to restart Claude for local plugin settings to take effect?

Yes, you need to restart Claude for local plugin settings to take effect safely. The configuration pattern requires documenting restart semantics so that updated YAML frontmatter values are loaded properly by hooks, commands, and agents.