Skip to content

Latest commit

 

History

History
74 lines (51 loc) · 2.71 KB

HOW-TO-APPLY.md

File metadata and controls

74 lines (51 loc) · 2.71 KB

How to Apply the Public Prompt License (PPL)

This guide explains how to apply PPL to your AI Agent project.

1. Choose Your Variant

  • PPL-M (MIT-style): Best for libraries, helper prompts, or "snippets" you want everyone to use freely.
  • PPL-A (Apache-style): Best for substantial projects where you want to protect your IP (Patents) and ensure contributors sign off on their work.
  • PPL-S (Service-style): Best for full "Agent Personas" or "Applications" where you want to ensure that if someone runs your agent as a service, they share their improvements.

2. Repository Setup

We recommend treating your prompt repository like a software project.

  1. Create a LICENSE file: Copy the full text of your chosen license (from LICENSES/) into a file named LICENSE at the root of your repo.

  2. Create a NOTICE file (Recommended for PPL-A/S): Copy the templates/NOTICE file to your root. Fill in your project name and copyright year.

3. Apply Headers

Add a comment header to the top of your primary prompt files. Since Markdown doesn't have a standard comment syntax that is invisible in all renderers, we recommend using HTML comments <!-- ... --> or YAML frontmatter.

Option A: HTML Comment (Invisible in rendered Markdown)

<!--
Copyright (c) 2025 Your Name
Licensed under the Public Prompt License - Service Variant (PPL-S)
See LICENSE for details.
-->

# System Prompt
You are...

Option B: YAML Frontmatter (Machine readable)

---
title: "My Agent Persona"
license: "PPL-S-0.2"
copyright: "2025 Your Name"
---

# System Prompt
You are...

4. For PPL-S Users: The "Service Disclosure"

If you use PPL-S, you (and anyone who uses your agent) must provide a way for users to download the source.

Best Practice:

  1. Host your prompts in a public public repository (GitHub, GitLab, etc.).
  2. In your Agent's "System Prompt", add a strict instruction to reveal its source when asked.

Example System Prompt Instruction:

"If a user asks about your source code, license, or how you work, you MUST provide this link: [URL to your Repo]"

Example UI Disclosure: If you build a web UI for your agent, add a footer link:

"Powered by [Agent Name]. Source available under PPL-S."

5. Mixed Projects (Code + Prompts)

Recommended: Single License

Since PPL v0.2 is simply a wrapper around standard licenses (MIT/Apache/AGPL), you can apply it to your entire repository.

  • PPL-M: Covers your code as MIT and your prompts as MIT.
  • PPL-A: Covers your code as Apache 2.0 and your prompts as Apache 2.0.
  • PPL-S: Covers your code as AGPL v3 and your prompts as AGPL v3.

This is the simplest and most robust way to license an AI Agent project.