Docs: how-to for extending Han skills via Claude plugin dependencies

[mxriverlynn]

Summary

Write a new how-to article (or articles, as needed) explaining how to extend Han skills using Claude plugin dependencies in plugin configuration. Source it from the canonical Claude docs and use han.github as the worked, in-repo example.

Background

han.github already extends Han the way this article should teach: it ships GitHub-facing skills and depends on han.core, and the han meta-plugin pulls both in through its dependencies. That makes it a concrete, shippable example of "extend Han via plugin dependencies" rather than a hypothetical.

Goal

A how-to that shows operators/contributors how to extend Han skills by depending on it from another plugin, with:

• Sourced information from canonical Claude documentation.
• Concrete examples pulled directly from han.github (its plugin.json dependency on han.core, its skills, and the han meta-plugin's dependencies).
• An explanation of how these pieces integrate and why they integrate that way.

Canonical reference

Start from the canonical Claude documentation on plugin dependencies:
https://code.claude.com/docs/en/plugin-dependencies#enable-or-disable-a-plugin-with-dependencies

Suspected areas / inputs

• han.github/ — the worked example (skills + han.github/.claude-plugin/plugin.json).
• han/.claude-plugin/plugin.json — the meta-plugin that depends on han.core + han.github.
• docs/how-to/ (workflow guide) vs. docs/guidance/ (contributor authoring guidance) — candidate homes.
• docs/guidance/claude-marketplace-and-plugin-configuration/ and docs/guidance/plugin-entity-taxonomy.md — existing references to build on, not duplicate.

Open questions to resolve before writing

• Audience: operators composing plugins, or contributors authoring a new dependent plugin? (Affects framing.)
• Scope of "extend": only the dependency-via-plugin.json mechanism, or also adding new skills/agents that build on existing Han skills?
• Placement & count: one article or a small set, and which folder.
• Done criteria: e.g., a reader can stand up a new plugin that depends on han.core and adds a skill.

Recommended next step

Run /research first — this is fundamentally a research-and-document task. The research report (canonical-doc findings + han.github evidence) becomes the source material for the article(s).

---

_{Triaged via /issue-triage. Type: Feature Request · Severity: Low.}