Browse all 9 articles in this open-source series. Nine evidence-led chapters on manifests, local distribution, overlay selection, draft hooks, service permissions and a read-only inspection project.
Featured starting point
Understand the openai/plugins repository, how packages combine skills and connections, and why its pinned compatibility layout differs from current portable guidance.
Suggested learning path
01 → 09
Understand the openai/plugins repository, how packages combine skills and connections, and why its pinned compatibility layout differs from current portable guidance.
Use the Figma example to trace manifest paths and connection assumptions, then define an isolated skills-only trial with a clear acceptance condition.
Separate package distribution from MCP service deployment and understand why editing a source folder may not update an installed local plugin.
Understand fixed portable component locations, the non-merging OpenAI overlay rule and why .app.json and MCP configuration represent different connections.
Trace the manifest, app mapping and hook script without mistaking a declared workflow or echoed reminder for an executed visual regression check.
Design an evidence-based plugin trial without assigning one price or speed claim to a heterogeneous catalogue of skills and connected services.
Review executable hooks and connected-service access separately from package metadata, and avoid assuming uninstalling a bundle revokes every connection.
Compare instruction reuse, package distribution and external-service tools without treating them as competing implementations of the same layer.
Design a read-only package-review tool that reports selected overlays, missing references and untested capabilities without installing or executing plugins.
