Reusable agent skills for the stacks we build on. The conventions we use as a company across the products we build, not generic best practices.
Each skill is a stack convention, a pattern you can drop into any project on that stack, not glue for one specific repo.
# Select skills interactively
npx skills add encurehq/skills| Skill | Stack | What it's for |
|---|---|---|
setup-agent-first |
Any stack | Bootstrap or retrofit a repo with AGENTS.md, structured docs, an ADR convention, and optional automation so AI agents can work without per-session hand-holding. Safe to re-run later — it detects its own prior setup and iterates (new docs/deps/patterns since last run) instead of starting over. |
elysia-pattern |
Elysia + Drizzle + Zod | Scaffold a backend module the canonical way: a Repository/Policy/Service trio, a routes file, and a one-line registration. |
elysia-zod |
Elysia + Zod | Decode 4xx validation errors fast — the six most common error shapes mapped to their exact fixes. |
next-monorepo-pattern |
Next.js monorepo | Where routes, components, hooks, and lib code live — the packages/ui → app → route priority chain, role-first forms and dialogs. |
next-tanstack-query |
Next.js + TanStack Query | The canonical read (lib/data, prefetch → hydration → useQuery) and write (lib/action, throws → toast, optimistic useMutation) data pattern. |
i18n-next-intl |
next-intl | Keep next-intl translations in sync across every locale. |
add-issues |
Any stack + Linear | Draft and publish one or more Linear issues, each a phased execution plan another agent can pick up cold and run — publishes directly, no confirmation step; falls back to chat/local files if the Linear MCP isn't connected. |