---
name: __APP_NAME__
description: This skill should be used when working on the __APP_NAME__ app — its workflows, services, schemas, config, deploys, and the conventions that are specific to it. Grow it as the app grows; it is the only skill that knows this app.
---

# __APP_NAME__

> **This file is yours.** `mesh create-app` seeded the stub; `mesh skills sync`
> never touches it again. It carries no managed marker on purpose — the app repo
> knows the app better than any platform skill can, so this is where the
> app-specific knowledge belongs.

## Where the platform knowledge lives (don't duplicate it here)

| For | Load |
|---|---|
| Scaffold → wire → run → debug a Mesh app | `.claude/skills/mesh-app-kit-apps/SKILL.md`, or `pnpm exec intent load @mesh-tech/app-kit#apps` |
| Exact `AppEnvironment` / primitive / Temporal API signatures | `pnpm exec intent load @mesh-tech/app-kit#core` |
| `mesh` CLI commands (`dev`, `deploy`, `stack`, `hub dev`, tunnels, db) | `.claude/skills/mesh-core/SKILL.md` |
| Everything installed that ships a skill | `pnpm exec intent list` |

Read those first. This file is for what they cannot know.

## What this app is

<!-- One paragraph: what __APP_NAME__ does and who calls it. -->

- **Path:** `__APP_PATH__`
- **Primitives:**
  <!-- e.g. service (Hono API), database (Prisma), temporal worker, bucket -->

## Conventions specific to this app

<!--
Fill these in as you learn them. Good candidates:
  - the domain vocabulary an agent will otherwise guess wrong
  - which module owns what, and what must never import what
  - the shape of the app's config and where each value comes from
  - required env/secrets and how they are provisioned
  - workflow/activity naming and versioning rules
  - test fixtures and how to run just this app's suite
-->

## Gotchas

<!--
Every time an agent (or you) gets something wrong twice, write it here.
This section is what makes the skill worth loading.
-->

## Runbook

```bash
cd __APP_PATH__
pnpm install
mesh dev            # run locally
mesh hub dev        # watch this session's runs in a local Hub
mesh deploy up      # deploy via the stack's deployer role
```