Contributing

Amplifiers is an open-source library of specialist skills, stacks, and skill-local supporting material.

It follows the Agent Skills package style used across the broader skills ecosystem, including OpenAI Codex.

The contribution rule is straightforward: put the material where the runtime will expect to find it, and keep the public model easy to understand.

Ways to contribute

• add a new skill under skills/<name>/SKILL.md
• add a new stack under stacks/<name>/README.md
• improve an existing skill
• improve a skill's references/, scripts/, or assets/
• improve repository documentation and maintainership guidance

Before opening a pull request

1. Keep each change focused.
2. Use clear English in docs and skill text.
3. Avoid introducing new architectural layers unless the use case is proven.
4. Read WRITING_STANDARD.md before adding new content or changing structure.
5. Confirm links still work.

Publishing flow

1. Start from the target skill package under skills/<name>/.
2. Publish a runtime skill only when one clear attachable specialist or modifier exists.
3. Move long supporting material into references/ when the runtime file gets too large.
4. Publish a stack when multiple skills form a recurring delivery pattern.
5. Update the public docs when the repository model or catalog changes.

Writing standard

Use WRITING_STANDARD.md for project-wide rules on:

• file naming
• SKILL.md versus README.md
• folder shapes for skills/, stacks/, and knowledge/
• tone and minimum quality expectations

If you are unsure where something belongs, ask this first:

• is this one attachable specialist
• is this a recurring combination of specialists
• is this deep supporting material for one skill
• is this public documentation about the repository itself