SHA256

Clement Verna fd63a60b89 docs: Reframe repo-level files as a skills library

Address review feedback on PR #2: the repo-level files were written
as if this is a single-skill project (Meeting Secretary) rather than
a library where Meeting Secretary is the first entry.

Changes:
- README.md: retitle as 'Fedora AI Skills Library', frame the repo
  as a library, reduce to library-level content, add Clement Verna
  to contributors
- AGENTS.md: reframe objective as building a skills library, drop
  meeting-secretary-specific details (those live in SKILL.md),
  describe the generic skills structure, update common tasks and
  layout table to library scope
- CONTRIBUTING.md: rewrite as library-level contribution guide
  (add a skill, improve a skill) rather than Meeting Secretary
  instructions
- CLAUDE.md: replace with a symlink to AGENTS.md

Assisted-by: Claude claude-sonnet-4-6@default

2026-05-26 19:30:09 +00:00

2.6 KiB

Raw Permalink Blame History

Contributing — Fedora AI Skills Library

Contributions are welcome — whether you are adding a new skill, refining an existing one, or improving the library documentation.

This project uses an agentic-first workflow: AI coding agents assist with drafting, while humans set intent, review, and own every change that ships.

Before you start

Read AGENTS.md — project context, architecture, and conventions.
Read REDHAT.md — Red Hat AI policy: attribution, sensitive data, upstream rules.

What you can contribute

Add a new skill

Create a new directory under skills/ named after your skill (lowercase, hyphens only).
Add a SKILL.md with YAML frontmatter (name, description) and instructions.
Optionally add references/ and assets/ subdirectories for supporting files.
Add your skill to the table in skills/README.md.
Submit a merge request.

The name field in frontmatter must match the directory name. See the Agent Skills specification for the full format.

Improve an existing skill

Each skill is self-contained in its own directory. To improve a skill's instructions, reference data, or templates:

Read the skill's SKILL.md to understand its current design.
Make your changes inside the skill directory.
Submit a merge request describing what you changed and why.

For substantial changes to a skill's core instructions, open an issue first to discuss the motivation and approach. For small fixes (typos, clarifications, outdated reference data), a direct merge request is fine.

Attribution

All AI-assisted contributions require a commit trailer per REDHAT.md:

Situation	Trailer
You directed the work and edited meaningfully	`Assisted-by: <tool name>`
Large generated chunk, minimal human edit	`Generated-by: <tool name>`
Unsure	`Assisted-by: <tool name>`

Code review expectations

Reviewers should focus on:

Accuracy — are the skill's instructions correct and grounded?
Scope — does the skill stay focused on its declared task?
Hallucination risk — could any change cause an agent to fabricate information?
Template consistency — do placeholders match what the skill instructions expect?

Conventions

Markdown: One sentence per line (ventilated prose).
Template placeholders: {{DOUBLE_BRACES}} syntax.
Skill naming: lowercase letters and hyphens only (e.g., meeting-secretary).
License: All contributions are made under CC BY-SA 4.0.

2.6 KiB Raw Permalink Blame History