
creating-plugins
Create EmDash CMS plugins with sandboxed hooks, routes, storage, content and media APIs, MCP tools, and declarative admin UI, or native React and Astro extensions. Use when scaffolding or implementing an EmDash plugin.
- Installs
- 128
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/emdash-cms/emdash --skill creating-plugins- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads creating-plugins automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Creating EmDash plugins
Build against the API that reaches the intended execution mode. Source types and production-boundary tests take precedence over examples in this skill when they disagree.
Choose a format
| Format | Runtime source | Admin UI | Distribution |
|---|---|---|---|
| Sandboxed | src/plugin.ts default-exports a SandboxedPlugin | Block Kit pages, widgets, saved-entry panels, and actions | Plugin CLI and registry |
| Native | definePlugin() / createPlugin() | React, Block Kit, and Astro components | Trusted site dependency only |
Use a sandboxed plugin unless the feature needs host-process access, React admin code, Astro rendering components, raw page fragments, or custom Portable Text block definitions. Native plugins run with the site's authority and cannot be installed from the registry.
Scaffold a sandboxed plugin
pnpm dlx @emdash-cms/plugin-cli init my-plugin
cd my-plugin
pnpm install
pnpm run test
The manifest is the identity and trust contract. Runtime hooks and routes live in src/plugin.ts; the CLI generates descriptors, manifests, and bundles. Do not create a separate descriptor factory or sandbox-entry.ts.
import type { SandboxedPlugin } from "emdash/plugin";
const plugin: SandboxedPlugin = {
hooks: {
"content:afterSave": async (event, ctx) => {
ctx.log.info("Content saved", { id: event.content.id });
},
},
};
export default plugin;
Import authoring types from emdash/plugin with import type. The CLI bundles value imports such as pluginRoute() and pluginResponse() from emdash/plugin and portableTextToMarkdown() from emdash/client. Sandboxed runtime code can use Web APIs but not Node.js built-ins.
Declare access
Declare every host API in emdash-plugin.jsonc. Adding authority, exposing a route publicly, or adding MCP tools requires renewed administrator approval.
| Capability | Grants |
|---|---|
schema:read | Public collection and field definitions |
admin.editor-draft:read | Selected unsaved field values after an explicit editor interaction |
admin.editor-draft:patch | Host-validated unsaved field changes proposed for editor review |
content:read | Content identity, translations, and published public URLs |
content:revisions:read | Retained revision data; implies content read |
content:write | Create, update, delete, and translation creation; implies read |
content:publish | Revision-fenced publish, unpublish, schedule, and unschedule; implies read |
content:restore | Revision-fenced reads and restoration of trashed content |
hooks.content-policy:register | Pre-publish, pre-schedule, and pre-unpublish policy hooks |
taxonomies:read | Taxonomy definitions, terms, and entry assignments |
taxonomies:write | Term creation and assignment deltas; implies read |
bylines:read | Public byline profiles and single or batched entry credits |
redirects:read | Versioned redirect inspection |
redirects:write | Versioned redirect creation, update, and deletion; implies read |
comments:read | Stored non-trashed comments and their personal data |
comments:moderate | Expected-status moderation; implies read |
media:read | Ready-media metadata and authenticated asset URLs |
media:bytes:read | Bounded media bytes and content hashes |
media:metadata:write | Alt text, caption, and focal-point updates |
media:write | Upload and delete; implies media read |
network:request | ctx.http.fetch() restricted to allowedHosts |
network:request:unrestricted | ctx.http.fetch() without a host list |
users:read | User directory lookup; also required by comment hooks |
email:send | Email delivery when a transport is configured |
hooks.email-transport:register | Exclusive email:deliver hook |
hooks.email-events:register | Email before/after hooks |
hooks.page-fragments:register | Trusted-only page fragments; excluded from sandbox registration |
Settings, KV, declared storage, logging, and cron scheduling are plugin-scoped and need no capability. Use ctx.settings for user configuration, ctx.kv for internal key-value state, and declared ctx.storage.<collection> for queryable records.
Read the focused reference for the API being used:
- Content, schema, translations, and publication
- Taxonomies and redirects
- Comments
- Media
- Storage, KV, and encrypted settings
Routes, HTTP, and MCP
Routes are private by default and require authentication, their declared RBAC permission, token scope or CSRF as appropriate. Public routes are internet-facing and require explicit install consent.
Undeclared routes keep the legacy JSON/query envelope. Declare methods and body modes for host-enforced parsing. Use pluginResponse() only on a route that declares a raw response. Outbound ctx.http.fetch() responses and declared route bodies are buffered and bounded.
MCP tools reference private JSON routes with explicit permissions and Zod input schemas. Mark difficult-to-reverse operations destructive: true. Read API routes and MCP tools.
Hooks
Sandboxed hooks enter the same priority, dependency, timeout, error-policy, enablement, exclusive-provider, and capability pipeline as trusted hooks. Publication policy hooks identify action origin and actor without granting publication authority. Comment moderation invoked through ctx.comments reports plugin origin and runs the normal after-hook once. Read Hooks.
Declarative admin UI
Sandboxed pages and widgets return validated Block Kit. Structured links use host-resolved targets; routeCtx.ui carries host-attested locale, direction, and surface. External images require matching network authority.
Saved-entry panels and actions point to private routes. Ordinary panel load receives only host-reloaded saved identity and version. Add admin.editor-draft:read or admin.editor-draft:patch plus extension-level collection and field selectors when an explicit interaction must receive selected unsaved fields or propose an atomic whole-field patch. Patch does not imply read. The host previews accepted patches, marks the form dirty, and never saves them automatically. Read saved content through capability-gated ctx.content.
Declarative field widgets currently compose supported Block Kit elements into a JSON value. Custom Portable Text blocks and Astro render components remain native-only. Read Admin UI, Block Kit, and Portable Text blocks.
Runner parity
Cloudflare Worker Loader and Node/workerd execute the same bundle behind a plugin-scoped bridge. Both return buffered WHATWG responses from ctx.http.fetch() with binary bytes preserved and decoded request/response bodies limited to 8 MiB. Write against exported types, not extra methods found in one wrapper.
Read Sandbox boundaries before designing around an API not listed here.
Test the production boundary
Use createPluginTestHost() for fast hook, route, manifest, capability, KV, settings, and storage transport tests. Use createPluginRuntimeTestHost() when the test must exercise real content actions, plugin activation, media, comments, redirects, scheduling, restart, authorization, CSRF, caching, Block Kit validation, or saved-entry extensions.
Runtime fixtures establish state without firing hooks. Runtime actions call production boundaries; inspectors read observable state. Queue outbound HTTP responses with host.http.respond() and inspect requests with host.http.requests(). Dispose every host after use. Add Node/workerd parity only for runner-sensitive behavior.
References
- Content, schema, translations, and publication
- Taxonomies and redirects
- Comments
- Media
- Hooks
- Storage, KV, and encrypted settings
- Admin UI and field widgets
- API routes and MCP tools
- Block Kit
- Portable Text blocks
- Sandbox boundaries
- Publishing
FAQ
What does the creating-plugins skill do?
Create EmDash CMS plugins with sandboxed hooks, routes, storage, content and media APIs, MCP tools, and declarative admin UI, or native React and Astro extensions. Use when scaffolding or implementing an EmDash plugin.
How do I install creating-plugins?
Run npx skills add https://github.com/emdash-cms/emdash --skill creating-plugins in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/creating-plugins/ (personal) or .claude/skills/creating-plugins/ (project).
Which AI agents can use creating-plugins?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is creating-plugins safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains creating-plugins?
It is published by emdash-cms in emdash-cms/emdash. It has 128 recorded installs.
More from emdash-cms/emdash
emdash-cli
emdash-cms/emdash
Use the EmDash CLI to inspect and manage an EmDash instance from the command line, including content, schema, media, taxonomies, menus, search, authentication, seeds, migrations, generated types, and whole-site export and import.
wordpress-plugin-to-emdash
emdash-cms/emdash
Open the skill to read its SKILL.md instructions.
wordpress-theme-to-emdash
emdash-cms/emdash
Open the skill to read its SKILL.md instructions.
agent-browser
emdash-cms/emdash
Open the skill to read its SKILL.md instructions.
Popular frontend & design skills
setup-matt-pocock-skills
mattpocock/skills
Configure this repo for the engineering skills: set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.
frontend-design
anthropics/skills
Guidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults.
hyperframes-animation
heygen-com/hyperframes
All animation knowledge for HyperFrames — atomic motion rules, multi-phase scene blueprints, scene transitions, broader motion-design techniques, AND the seven runtime adapters (GSAP default, plus Lottie, Three.js, Anime.js, CSS keyframes, Web Animations API, TypeGPU). Use for any motion or animation task: pick 2-4 rules and compose, or load a blueprint, or look up runtime-specific API (e.g. GSAP eases / Lottie player / Three.js mixer). Also covers auditing an existing composition's choreography (animation map) and 24 named text-animation effects. HyperFrames-native: single paused timeline, seek-safe, deterministic.
prototype
mattpocock/skills
Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.
hyperframes-core
heygen-com/hyperframes
The HyperFrames composition contract — build one renderable project. Use for composition structure, the data- timing attributes, class="clip", tracks, sub-compositions, variables, framework-owned media playback, deterministic-render rules, and validation. Read before writing composition HTML.
vercel-react-best-practices
vercel-labs/agent-skills
React and Next.js performance optimization guidelines from Vercel Engineering. This skill should be used when writing, reviewing, or refactoring React/Next.js code to ensure optimal performance patterns. Triggers on tasks involving React components, Next.js pages, data fetching, bundle optimization, or performance improvements.