NO-JIRA: Add a dev container setup and a basic MCP server

.apm/instructions/dev-commands.instructions.md

@@ -0,0 +1,22 @@
+---
+description: "Common dev commands for database migration, linting, and testing"
+applyTo: "**"
+---
+
+### Database migration
+
+Run migrations: `go run ./cmd/sippy migrate --database-dsn $SIPPY_DATABASE_DSN`
+
+If `SIPPY_DATABASE_DSN` is not set, use the dev default: `postgresql://postgres:password@localhost:5432/postgres`
+
+### Linting
+
+Run lint: `CI=true make lint`
+
+`CI=true` makes `hack/go-lint.sh` use the locally installed `golangci-lint` instead of spawning a container.
+
+### Testing
+
+Run unit tests: `make test`
+
+This runs Go tests via gotestsum and sippy-ng Jest tests.

.apm/instructions/mcp.instructions.md

@@ -0,0 +1,8 @@
+---
+description: "MCP server (sippy-dev) for AI-callable dev tasks"
+applyTo: "mcp/**"
+---
+
+Shared MCP server for AI-callable dev tasks (migrate, serve, lint, test, e2e). Configuration, tool list, logs, and extension notes: **[README.md](../../mcp/README.md)**.
+
+When adding or modifying MCP tools, follow existing patterns in `server.py` (subprocess, `_repo_path`, `_ensure_dev_log_dir`, credentials helpers). Restart the MCP server after changes.

.apm/prompts/sippy-dev-app.prompt.md

@@ -0,0 +1,12 @@
+---
+description: "Start both Sippy backend and frontend dev servers"
+---
+
+# Sippy dev — backend + frontend
+
+Start the full local Sippy stack using two MCP tool calls (server: **`sippy-dev`**). Run them in order — backend first, then frontend.
+
+1. **`sippy_serve`** — pass **`bigquery_credentials_file`** when `SIPPY_BIGQUERY_CREDENTIALS_FILE` / `GOOGLE_APPLICATION_CREDENTIALS` are not set.
+2. **`sippy_ng_start`**
+
+Each tool returns listen hints (typically **8080** / **3000**) and log paths. If a tool reports already running, ask the user if they want to restart it. If yes, call the tool again with **`restart=True`**.

.apm/prompts/sippy-dev-frontend.prompt.md

@@ -0,0 +1,13 @@
+---
+description: "Start the sippy-ng React dev server via the sippy-dev MCP tool"
+---
+
+# Sippy dev — frontend (sippy-ng)
+
+Use the **`sippy_ng_start`** MCP tool (server: **`sippy-dev`**). Do not run `npm start` in `sippy-ng` manually — the MCP tool handles background process management, log routing, and duplicate detection.
+
+**`open_browser`** defaults to **`false`**. Typical URL: **`http://127.0.0.1:3000`**. Log: **`sippy-dev-logs/sippy_ng_start.log`**.
+
+If the dev server is already running, the tool will report it. Ask the user if they want to restart, and if so call again with **`restart=True`**.
+
+See **`mcp/server.py`** for all parameters.

.apm/prompts/sippy-dev-migrate.prompt.md

@@ -0,0 +1,17 @@
+---
+description: "Run Sippy PostgreSQL schema migration"
+---
+
+# Sippy dev — migrate
+
+Run the migration command directly:
+
+```bash
+go run ./cmd/sippy migrate --database-dsn "$SIPPY_DATABASE_DSN"
+```
+
+If `SIPPY_DATABASE_DSN` is not set, use the dev default:
+
+```bash
+go run ./cmd/sippy migrate --database-dsn "postgresql://postgres:password@localhost:5432/postgres"
+```

.apm/prompts/sippy-dev-regression-cache.prompt.md

@@ -0,0 +1,11 @@
+---
+description: "Run the Sippy regression-cache loader (BigQuery + Redis + DB)"
+---
+
+# Sippy dev — regression-cache
+
+Use the **`regression_cache`** MCP tool (server: **`sippy-dev`**). Do not run `go run ./cmd/sippy load --loader regression-cache` manually — the MCP tool handles credentials, logging, and timeouts.
+
+**`bigquery_credentials_file`**: path to BigQuery-capable SA JSON (e.g. `sippy-bigquery-job-importer-key.json`); optional if `SIPPY_BIGQUERY_CREDENTIALS_FILE` or `GOOGLE_APPLICATION_CREDENTIALS` is set.
+
+See **`mcp/server.py`** for all parameters. Log: **`sippy-dev-logs/regression_cache.log`**. Typical duration is many minutes.

.apm/prompts/sippy-dev-serve.prompt.md

@@ -0,0 +1,13 @@
+---
+description: "Start the Sippy HTTP API server via the sippy-dev MCP tool"
+---
+
+# Sippy dev — serve
+
+Use the **`sippy_serve`** MCP tool (server: **`sippy-dev`**). Do not run `go run ./cmd/sippy serve` manually — the MCP tool handles background process management, log routing, and duplicate detection.
+
+**`bigquery_credentials_file`**: path to BigQuery-capable SA JSON (e.g. `sippy-bigquery-job-importer-key.json`); optional if `SIPPY_BIGQUERY_CREDENTIALS_FILE` or `GOOGLE_APPLICATION_CREDENTIALS` is set.
+
+If the server is already running, the tool will report it. Ask the user if they want to restart, and if so call again with **`restart=True`**.
+
+See **`mcp/server.py`** for all parameters. Typical listen address: **`:8080`**. Log: **`sippy-dev-logs/sippy_serve.log`**.

.apm/prompts/sippy-dev-setup.prompt.md

@@ -0,0 +1,113 @@
+---
+description: "Set up the Sippy devcontainer (Podman, env, GCP auth)"
+---
+
+# Sippy dev — devcontainer setup
+
+Interactive setup for the Sippy devcontainer. Automates what can be detected, prompts for what can't.
+
+## Workflow
+
+### 1. Detect OS
+
+```bash
+uname -s
+```
+
+- **Darwin** = macOS
+- **Linux** = Linux
+
+### 2. Check prerequisites
+
+Verify each tool is installed. Report any that are missing and stop.
+
+```bash
+command -v podman
+command -v devcontainer
+```
+
+If `devcontainer` is missing, tell the user: `npm install -g @devcontainers/cli`
+
+### 3. macOS: Podman machine
+
+If macOS, check if podman machine is running:
+
+```bash
+podman machine info
+```
+
+If not initialized or not running, run:
+
+```bash
+podman machine init   # only if no machine exists
+podman machine start
+```
+
+### 4. Linux: Podman socket
+
+If Linux, check if the socket is active:
+
+```bash
+systemctl --user is-active podman.socket
+```
+
+If not active, run:
+
+```bash
+systemctl --user enable --now podman.socket
+```
+
+Also check for `podman-docker`:
+
+```bash
+command -v docker
+```
+
+If missing, suggest installing `podman-docker` (dnf or apt depending on `/etc/os-release`).
+
+### 5. Environment file
+
+Check if `.devcontainer/.env` exists:
+
+```bash
+test -f .devcontainer/.env
+```
+
+If missing, copy from the example:
+
+```bash
+cp .devcontainer/.env.example .devcontainer/.env
+```
+
+Then read `.devcontainer/.env` and check for empty required values. If any are blank (e.g. `ANTHROPIC_VERTEX_PROJECT_ID`), tell the user which values need to be filled in and ask them to edit `.devcontainer/.env` and let you know when they're done. **Do not** ask for the values directly or write to the file yourself — the user should edit it. Wait for them to confirm before continuing.
+
+### 6. Start the container
+
+Determine the right command based on OS:
+
+- **macOS**: `devcontainer up --workspace-folder . --docker-path podman`
+- **Linux** (with `podman-docker`): `devcontainer up --workspace-folder .`
+- **Linux** (without `podman-docker`): `devcontainer up --workspace-folder . --docker-path podman`
+
+Run it. This triggers `init-services.sh` (PostgreSQL + Redis) and `post-create.sh` (Go tools, npm, MCP venv).
+
+### 7. GCP authentication
+
+GCP credentials are mounted from the host's `~/.config/gcloud` directory. If the user hasn't authenticated on the host yet, tell them to run on the host (not inside the container):
+
+```bash
+gcloud auth application-default login
+```
+
+The credentials will be available inside the container automatically after restart.
+
+### 8. Summary
+
+Print a summary of what was set up:
+
+- Container status
+- PostgreSQL: `localhost:5432`
+- Redis: `localhost:6379`
+- API server: `localhost:8080` (start with `/sippy-dev-serve`)
+- React dev server: `localhost:3000` (start with `/sippy-dev-frontend`)
+- GCP auth status

.apm/prompts/sippy-dev-tests.prompt.md

@@ -0,0 +1,25 @@
+---
+description: "Run lint, unit tests, and e2e in order (full local CI suite)"
+---
+
+# Sippy dev — full test suite
+
+Run these three steps in order. Stop if any step fails.
+
+1. **Lint** — run directly:
+
+   ```bash
+   CI=true make lint
+   ```
+
+   `CI=true` makes `hack/go-lint.sh` use the locally installed `golangci-lint` instead of spawning a container.
+
+2. **Unit tests** — run directly:
+
+   ```bash
+   make test
+   ```
+
+3. **E2e** — use the **`run_e2e`** MCP tool (server: **`sippy-dev`**). Pass **`bigquery_credentials_file`** when cred env vars are unset. E2e requires Podman/Docker and **does not work inside the devcontainer** — run it on the host.
+
+E2e log: **`sippy-dev-logs/run_e2e.log`**. Timeouts: **`mcp/server.py`**.