Contributing

The toolkit grows by FDE contribution. Most additions are new skills layered onto existing plugins; occasional larger contributions add a whole new plugin.

Common contribution paths

What you’re adding	Start here
A new slash command or workflow	Adding a Skill
A new always-on rule (e.g., a project convention)	Adding a Skill — same flow, drop into rules/ instead of skills/
A new specialist agent	Adding a Skill — drop into agents/
A whole new plugin (e.g., a vertical-specific toolkit)	Adding a Plugin

Litmus tests before contributing

Skill or rule?

Does wrapping the guidance in a skill change Claude’s output compared to just asking the same question directly?

• Yes → ship as skills/<name>/SKILL.md
• No → ship as rules/<topic>.md (always-on context)

Skill or slash command?

• Skill auto-matches from conversation context. Good for workflows that should activate when relevant.
• Slash command is user-invoked with /command. Good for explicit operations with arguments.
• Often a workflow deserves both — a skill defines the knowledge and steps; a slash command is the invocation handle.

Validation before PR

Run claude plugin validate against the plugin you changed plus the marketplace manifest, and smoke-test the change in a fresh session. Full flow: Validation Workflow.

What NOT to add

These are out of scope for the toolkit. They belong on the FDE workstation as user-scope assets, not in the marketplace:

• Format-on-save hooks — conflicts with team Prettier/ESLint configs
• Auto-fix skills that skip human review
• Client-specific assets, templates, or brand content — stays local on the FDE workstation
• FDE-personal meeting notes, prior conversations, or research — lives in ~/.claude/skills/ (user scope), not in this repo

Source

The canonical contributor guide is CONTRIBUTING.md in the repo root.