# Plugin Best Practices

> Agent-plugin design conventions for public outcomes, private implementation, and minimal packaging. Use when creating, restructuring, or reviewing agent-skill plugins.

- Skill: `bastidood/plugin-best-practices` (Agent Skill, multi-file: 4 files)
- Install (CLI): `npx skillmds@latest add bastidood/plugin-best-practices`
- Raw SKILL.md: https://api.skillmd.com/api/skills/bastidood/plugin-best-practices/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: BastiDood (https://skillmd.com/u/bastidood)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/bastidood/plugin-best-practices

---


# Plugin Best Practices

Prefer fine-grained plugins that package one coherent installable capability or body of guidance. Consumers install the plugin as a unit instead of assembling cooperating skills or accepting unrelated functionality.

Treat a plugin as an encapsulation boundary. Like a class in object-oriented programming, it exposes a deliberate public interface while hiding implementation details that consumers do not need to understand or invoke.

A plugin can expose one public skill or several independently useful public skills. Those entry points own consumer-facing outcomes while private skills support them without exposing the internal call graph.

## Effective Strategies for Focused Agent-Skill Plugins

Follow the guidelines below. Read each linked reference that applies to the plugin design task.

1. Expose public outcomes while keeping implementation skills private.
   - [Structure the plugin around public outcomes and private implementation skills.](./references/plugin-structure.md)
2. Package only the metadata and files that consumers need.
   - [Keep manifests and packaged files minimal.](./references/plugin-manifests.md)