Minn Admin exists because the WordPress admin buries everyday work under twenty years of accumulated chrome. These are the principles that decide what gets built and how. The measurable outcomes these principles serve — and the numbers that judge them — live in roadmap.md.
- Get out of the way. The daily work — writing, moderating, checking on the site, managing files and users — should be one click away and visually calm. Density serves the reader, not the toolbar.
- Real replacement, not a demo. Minn must hold up on real sites with hundreds of thousands
of rows (it's developed against a production-scale dataset). Pagination,
_fieldsallowlists and capability gating are non-negotiable, not optimizations. - Zero lock-in. Everything Minn writes is native WordPress: Gutenberg block markup, core options, core REST calls. Deactivate the plugin and nothing is lost or broken. Classic wp-admin remains fully available at all times — Minn is additive.
- No build step. Browser-native JavaScript, one stylesheet and PHP run as shipped. A contributor can trace behavior without reconstructing generated bundles. This is a promise about inspectability, not about keeping the whole codebase small enough to read in one sitting.
- Defensive by architecture. Never run other plugins' render paths in list contexts (a misbehaving plugin must not be able to take Minn down), never instantiate arbitrary objects from foreign blobs, prefer bounded parsers and use constrained decoding only when a vendor's storage or API requires it, check capabilities server-side, escape everything at the edge.
- An ecosystem invitation. Third-party plugins integrate through one declarative filter —
see
for-plugin-authors.md. Minn ships adapters for popular plugins; plugin authors can ship their own without writing JavaScript. - The user outranks every integration. Nothing a plugin registers may grab attention the
user did not ask for, and anything a plugin registers can be hidden or muted by the user.
WordPress lost this fight in its notification system; Minn enforces it with architecture
(descriptors only, extraction not hosting), budgets and per-user controls — never with
trust in plugin authors' restraint. See
v1-readiness.md.
- Gutenberg parity. Complex block layouts belong in the block editor; Minn links to it
cleanly (
editor-direction.md). - Settings-page parity. Minn surfaces the settings people actually change. The long tail stays in wp-admin.
- Network Admin parity. Minn covers daily multisite operations. Network account creation and deletion, the long settings tail, restores, exports and network setup stay in WordPress Network Admin.
- Being a page builder.
Every feature lands with: browser-level verification on a clean site and a production-scale site, zero console errors, capability checks proven by the UI hiding/showing correctly, and an Emoji-Log commit. If a feature can't be verified end-to-end, it isn't done.
