@dk/jolly 0.1.0

package/.env.example

@@ -0,0 +1,3 @@
+# Saleor Cloud API Token
+# Get your token at: https://cloud.saleor.io/settings/api-tokens
+SALEOR_CLOUD_TOKEN=your-token-here

package/.mcp.json

@@ -0,0 +1,7 @@
+{
+  "mcpServers": {
+    "saleor": {
+      "url": "https://mcp.saleor.app"
+    }
+  }
+}

package/.sisyphus/boulder.json

@@ -0,0 +1,13 @@
+{
+  "active_plan": "/var/home/hands/Work/jolly/.sisyphus/plans/saleor-agent-cli.md",
+  "started_at": "2026-04-02T12:03:12.876Z",
+  "session_ids": [
+    "ses_2b27f2227ffekotd1vwCAlGQKE",
+    "ses_2b0878847ffe4HYHHUEZz4sU2v",
+    "ses_2b079461affeWdGy097VEA4gGr",
+    "ses_2b06fe52cffesH6qKzJwV0pSn1"
+  ],
+  "plan_name": "saleor-agent-cli",
+  "agent": "atlas",
+  "task_sessions": {}
+}

package/.sisyphus/notepads/saleor-agent-cli/decisions.md

@@ -0,0 +1,11 @@
+# Decisions - Jolly CLI
+
+## Package Structure
+- Single package @saleor/jolly with multiple bin entries
+- Bun runtime, TypeScript throughout
+- OpenTUI for rich terminal output
+
+## CLI Architecture
+- yargs for argument parsing
+- Command-based structure (store, app, agent, config)
+- Exit codes: 0 success, 1 error

package/.sisyphus/notepads/saleor-agent-cli/issues.md

@@ -0,0 +1,6 @@
+# Issues - Jolly CLI
+
+## OpenTUI Consideration
+- OpenTUI requires Zig to build
+- May need fallback to simpler TUI if Zig unavailable
+- Check @opentui/core availability on npm

package/.sisyphus/notepads/saleor-agent-cli/learnings.md

@@ -0,0 +1,6 @@
+# Learnings - Jolly CLI
+
+## Project Setup
+- Bun runtime for TypeScript CLI
+- @saleor/jolly scoped package name
+- Three entry points: jolly, create-saleor-jolly, init-saleor-jolly

package/.sisyphus/plans/saleor-agent-cli.md

@@ -0,0 +1,600 @@
+# Jolly - Saleor Project Bootstrapper & Agent Configurator
+
+## TL;DR
+
+> **Quick Summary**: Jolly bootstraps Saleor projects and configures local AI agents (OpenCode, Claude Code) with Saleor skills, rules, and MCP access — enabling agents to work with Saleor out of the box.
+> 
+> **Deliverables**:
+> - `jolly` CLI with `bootstrap` command (creates Cloud + Storefront + Payment App)
+> - `jolly setup-agents` command (configures OpenCode/Claude Code with skills, rules, MCP)
+> - Rich TUI output via OpenTUI
+> 
+> **Estimated Effort**: Small
+> **Parallel Execution**: YES - 2 waves
+> **Critical Path**: Project scaffolding → Bootstrap → Agent setup
+
+---
+
+## Context
+
+### Original Request
+Build a CLI for agents to use Saleor, inspired by swamp.club. Should:
+- Install Saleor's agent skills
+- Bootstrap complete new Saleor projects using saleor.io (Saleor Cloud)
+- Be TypeScript with Bun runtime
+
+### Clarified Vision
+**Jolly replaces saleor CLI for the agent age**. It enables agents to work with Saleor:
+1. **Project Bootstrapper** - Creates new Saleor projects (Cloud + Storefront + Payment App)
+2. **Agent Configurator** - Sets up local AI agents with Saleor capabilities
+3. **Agent Empowerment** - Agents can do everything they could do with manual CLI **and more**
+
+Jolly **provides equivalent functionality** to saleor cli via:
+- Saleor Cloud API (for all project/environment operations)
+- Official saleor-mcp (for agent tool access to store data)
+- saleor/agent-skills (for domain knowledge and best practices)
+
+The agent using jolly + saleor-mcp + skills can:
+- Query stores (products, orders, customers)
+- Manage environments
+- Deploy configurations
+- **Plus**: Agent reasoning, automation, and enhanced capabilities
+
+### Interview Summary
+**Key Decisions**:
+- **Bootstrap scope**: Saleor Cloud env + paper Storefront + hosted Payment App (Dummy/Stripe)
+- **Agent setup**: Configure OpenCode + Claude Code with skills, rules, MCP
+- **Stack**: Bun runtime, TypeScript, OpenTUI for rich TUI, direct Saleor API calls
+- **Test strategy**: Bun's built-in test runner with BDD-inspired architecture
+
+### Research Findings
+- **Saleor Cloud API**: Direct API for creating environments, registering apps
+- **paper Storefront** (`saleor/storefront`): React-based storefront template
+- **Payment Apps**: Hosted on saleor.io - Dummy and Stripe options
+- **Official Saleor MCP** (`mcp.saleor.app`): MCP server with channels, customers, orders, products, stocks
+- **saleor/agent-skills**: Portable skills (saleor-app, saleor-configurator, saleor-core, saleor-storefront)
+- **OpenTUI** (`@opentui/core`): Native Zig TUI core with TypeScript bindings
+
+---
+
+## Work Objectives
+
+### Core Objective
+Build a CLI tool that **creates new Saleor ecosystem components** and configures agents:
+1. **New Stores** - Create Saleor Cloud environments
+2. **New Apps** - Scaffold Saleor apps (Dashboards, Payment, Integrations)
+3. **New Agents** - Configure AI agents to work with Saleor
+4. **Configurator Integration** - Deploy/store configuration as code
+
+### Concrete Deliverables
+- `jolly` CLI executable (Bun-based)
+- **Store commands** - Create/manage Saleor Cloud stores
+- **App commands** - Scaffold new Saleor apps (Dashboard extensions, Payment apps, webhooks)
+- **Agent commands** - Configure AI agents with Saleor capabilities
+- **Config commands** - Deploy/configure stores via Configurator
+
+### Definition of Done
+- [ ] `jolly store create --name my-store` creates Saleor Cloud store
+- [ ] `jolly app create --name my-app --type dashboard-extension` scaffolds app
+- [ ] `jolly agent setup` configures agent with skills + AGENTS.md + MCP
+- [ ] All commands have proper error handling and exit codes
+- [ ] BDD tests pass for core functionality
+
+### Must Have
+- **Store creation**: Create Saleor Cloud environments
+- **App scaffolding**: New Dashboard extensions, Payment apps, webhook integrations
+- **Agent setup**: Configure OpenCode/Claude Code with skills, rules, MCP
+- **Config integration**: Deploy/configure stores via Configurator
+- Rich TUI output via OpenTUI
+
+### Must NOT Have (Guardrails)
+- No Saleor CLI wrapping (jolly uses APIs directly, doesn't wrap CLI)
+- No local payment app development (uses hosted apps only)
+- No web UI or dashboard
+- No plugin system in v1
+- No telemetry without consent
+
+---
+
+## Verification Strategy
+
+### Test Decision
+- **Infrastructure exists**: NO - New project
+- **Automated tests**: YES - BDD with Bun test
+- **Framework**: Bun's built-in test runner
+- **Style**: BDD-inspired (describe/it + Given/When/Then helpers)
+
+### QA Policy
+Every task includes agent-executed QA scenarios:
+- **CLI**: Use interactive_bash (tmux) to run commands and verify output
+- **API**: Use Bash (curl) for Saleor Cloud API calls
+- **Agent Config**: Verify files created and formatted correctly
+
+---
+
+## Execution Strategy
+
+### npm Distribution Entry Points
+
+| Command | Entry | Function |
+|---------|-------|----------|
+| `npx @saleor/jolly` | `jolly` bin | Full CLI (all commands) |
+| `npm create @saleor/jolly` | `create-saleor-jolly` bin | Store bootstrap (interactive) |
+| `npm init @saleor/jolly` | `init-saleor-jolly` bin | Agent setup |
+
+### CLI Commands (via `npx @saleor/jolly`)
+
+```
+jolly store create --name <name>           # Create Saleor Cloud store
+jolly store list                            # List your stores
+jolly store env create --store <id>        # Create environment
+
+jolly app create --name <name> --type <type>  # Scaffold new app
+  # Types: dashboard-extension, payment-app, webhook-handler
+
+jolly agent setup                          # Detect IDE, configure all (skills + AGENTS.md + MCP)
+jolly agent skills install                  # Install saleor/agent-skills (IDE-aware paths)
+
+jolly config deploy --store <id>          # Deploy via Configurator
+jolly config introspect --store <id>       # Introspect current config
+```
+
+### Parallel Execution Waves
+
+```
+Wave 1 (Foundation):
+├── Task 1: Project scaffolding + package.json + tsconfig
+├── Task 2: Core CLI structure (OpenTUI + yargs)
+├── Task 3: Saleor Cloud API client
+├── Task 4: OpenTUI components for output
+└── Task 5: BDD test setup
+
+Wave 2 (Store & App Commands):
+├── Task 6: Store commands - create/list environments
+├── Task 7: App scaffold command - dashboard extensions
+├── Task 8: App scaffold command - payment apps (hosted)
+├── Task 9: App scaffold command - webhook handlers
+
+Wave 3 (Agent Commands):
+├── Task 10: Agent setup command - skills installation
+├── Task 11: Agent setup command - AGENTS.md generation
+├── Task 12: Agent setup command - MCP configuration
+└── Task 13: Integration tests
+
+Wave FINAL (Verification):
+├── Task F1: Plan compliance audit
+├── Task F2: Code quality review
+├── Task F3: Real Manual QA
+└── Task F4: Scope fidelity check
+```
+
+---
+
+## TODOs
+
+- [ ] 1. Project scaffolding + package.json + tsconfig
+
+  **What to do**:
+  - Initialize Bun TypeScript project
+  - Set up tsconfig.json for CLI build
+  - Configure bun build for executable
+  - Add dependencies: yargs, @opentui/core, @saleor/configurator
+  - Set up package.json with bin entries for npm distribution:
+
+  ```json
+  {
+    "name": "@saleor/jolly",
+    "bin": {
+      "jolly": "./dist/cli.js",
+      "create-saleor-jolly": "./dist/bootstrap.js",
+      "init-saleor-jolly": "./dist/agent.js"
+    }
+  }
+  ```
+
+  **npm Entry Points**:
+  - `npx @saleor/jolly` → `jolly` bin (full CLI)
+  - `npm create @saleor/jolly` → `create-saleor-jolly` bin (store bootstrap)
+  - `npm init @saleor/jolly` → `init-saleor-jolly` bin (agent setup)
+
+  **References**:
+  - `https://opentui.com` - OpenTUI framework
+  - `https://github.com/anomalyco/opentui` - OpenTUI GitHub
+  - npm bin naming convention for scoped packages
+
+  **QA Scenarios**:
+  ```
+  Scenario: Build produces working executable
+    Tool: interactive_bash
+    Steps:
+      1. bun run build
+      2. ./dist/jolly.js --version
+    Expected: Version string printed
+    Evidence: .sisyphus/evidence/task-1-build.{ext}
+
+  Scenario: CLI help displays correctly
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js --help
+    Expected: Help text with all commands listed
+    Evidence: .sisyphus/evidence/task-1-help.{ext}
+  ```
+
+- [ ] 2. Core CLI structure (OpenTUI + yargs)
+
+  **What to do**:
+  - Set up yargs for argument parsing
+  - Integrate OpenTUI renderer for rich output
+  - Create base CLI with styled help/version
+  - Error handling with styled displays
+  - Exit codes (0 success, 1 error)
+
+  **QA Scenarios**:
+  ```
+  Scenario: CLI returns exit code 0 on success
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js --version
+    Expected: Exit code 0
+    Evidence: .sisyphus/evidence/task-2-exit-success.{ext}
+
+  Scenario: CLI returns exit code 1 on error
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js invalid-command
+    Expected: Exit code 1 with error message
+    Evidence: .sisyphus/evidence/task-2-exit-error.{ext}
+  ```
+
+- [ ] 3. Saleor Cloud API client
+
+  **What to do**:
+  - Create HTTP client for Saleor Cloud API
+  - Auth via SALEOR_CLOUD_TOKEN env var
+  - Endpoints: create environment, register app, list apps
+
+  **References**:
+  - Saleor Cloud API docs (saleor.io/cloud)
+  - API base URL: https://cloud.saleor.io/api/
+
+  **QA Scenarios**:
+  ```
+  Scenario: API client handles missing token gracefully
+    Tool: interactive_bash
+    Steps:
+      1. SALEOR_CLOUD_TOKEN="" ./dist/jolly.js store list
+    Expected: Error message about missing token, exit code 1
+    Evidence: .sisyphus/evidence/task-3-missing-token.{ext}
+
+  Scenario: API client creates store (mocked)
+    Tool: interactive_bash
+    Steps:
+      1. SALEOR_CLOUD_TOKEN=test-mock ./dist/jolly.js store create --name test
+    Expected: Mock response or appropriate error handling
+    Evidence: .sisyphus/evidence/task-3-store-create.{ext}
+  ```
+
+- [ ] 4. OpenTUI components for output
+
+  **What to do**:
+  - Create reusable OpenTUI components: Box, Text, Spinner, Progress
+  - Style system matching OpenCode aesthetic (dark theme, syntax colors)
+  - Consistent formatting for all CLI output
+
+  **QA Scenarios**:
+  ```
+  Scenario: OpenTUI Text component renders with styling
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js --help
+    Expected: Styled text output (not plain)
+    Evidence: .sisyphus/evidence/task-4-text-output.{ext}
+
+  Scenario: OpenTUI Spinner shows during operation
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js store create --name test  # with spinner
+    Expected: Spinner animation visible during operation
+    Evidence: .sisyphus/evidence/task-4-spinner.{ext}
+  ```
+
+- [ ] 5. BDD test setup
+
+  **What to do**:
+  - Set up Bun test with BDD describe/it
+  - Add Given/When/Then helpers
+  - Basic CLI structure tests
+
+  **QA Scenarios**:
+  ```
+  Scenario: BDD tests run successfully
+    Tool: interactive_bash
+    Steps:
+      1. bun test
+    Expected: All tests pass (0 failures)
+    Evidence: .sisyphus/evidence/task-5-bdd-run.{ext}
+
+  Scenario: Given/When/Then helpers work
+    Tool: interactive_bash
+    Steps:
+      1. bun test src/test/helpers.test.ts
+    Expected: BDD helper tests pass
+    Evidence: .sisyphus/evidence/task-5-bdd-helpers.{ext}
+  ```
+
+- [ ] 6. Store commands - create/list environments
+
+  **What to do**:
+  - `jolly store create --name <name>` - Create Saleor Cloud store
+  - `jolly store list` - List your stores
+  - `jolly store env create` - Create environment
+  - Call Saleor Cloud API directly
+
+  **QA Scenarios**:
+  ```
+  Scenario: store create command creates store (mocked)
+    Tool: interactive_bash
+    Steps:
+      1. SALEOR_CLOUD_TOKEN=test ./dist/jolly.js store create --name test-store
+    Expected: Store creation initiated or mock response
+    Evidence: .sisyphus/evidence/task-6-store-create.{ext}
+
+  Scenario: store list command lists stores
+    Tool: interactive_bash
+    Steps:
+      1. SALEOR_CLOUD_TOKEN=test ./dist/jolly.js store list
+    Expected: List output (may be empty)
+    Evidence: .sisyphus/evidence/task-6-store-list.{ext}
+  ```
+
+- [ ] 7. App scaffold command - Dashboard extensions
+
+  **What to do**:
+  - `jolly app create --name <name> --type dashboard-extension`
+  - Scaffold new Dashboard extension app
+  - Register with Saleor Cloud
+
+  **References**:
+  - `https://github.com/saleor/saleor-app-sdk` - App SDK
+  - Dashboard extension structure: src/pages/add/, useAppBridge()
+
+  **QA Scenarios**:
+  ```
+  Scenario: app create scaffold dashboard-extension
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js app create --name my-dashboard --type dashboard-extension
+    Expected: Directory my-dashboard created with dashboard extension template
+    Evidence: .sisyphus/evidence/task-7-dashboard-scaffold.{ext}
+
+  Scenario: app create shows help for missing name
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js app create --type dashboard-extension
+    Expected: Error about missing required --name flag
+    Evidence: .sisyphus/evidence/task-7-missing-name.{ext}
+  ```
+
+- [ ] 8. App scaffold command - Payment apps (hosted)
+
+  **What to do**:
+  - `jolly app create --name <name> --type payment`
+  - Options: Dummy Payment App OR Stripe Payment App (hosted on saleor.io)
+  - Register hosted payment app with Saleor Cloud
+
+  **QA Scenarios**:
+  ```
+  Scenario: app create scaffold payment with default (dummy)
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js app create --name my-payment --type payment
+    Expected: Payment app registered with Dummy (default)
+    Evidence: .sisyphus/evidence/task-8-payment-dummy.{ext}
+
+  Scenario: app create scaffold payment with stripe
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js app create --name my-payment --type payment --provider stripe
+    Expected: Payment app registered with Stripe
+    Evidence: .sisyphus/evidence/task-8-payment-stripe.{ext}
+  ```
+
+- [ ] 9. App scaffold command - Webhook handlers
+
+  **What to do**:
+  - `jolly app create --name <name> --type webhook`
+  - Scaffold webhook handler app
+  - Register with Saleor Cloud
+
+  **QA Scenarios**:
+  ```
+  Scenario: app create scaffold webhook
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js app create --name my-webhook --type webhook
+    Expected: Directory my-webhook created with webhook handler template
+    Evidence: .sisyphus/evidence/task-9-webhook-scaffold.{ext}
+
+  Scenario: webhook app has correct structure
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js app create --name test-webhook --type webhook
+      2. ls test-webhook/src/webhooks/
+    Expected: Webhook handler files present
+    Evidence: .sisyphus/evidence/task-9-webhook-structure.{ext}
+  ```
+
+- [ ] 10. Agent setup - Skills installation (IDE-aware)
+
+  **What to do**:
+  - `jolly agent skills install`
+  - Follow agentskills.io best practices for skill structure
+  - Detect available agents: OpenCode, Claude Code, OpenClaw, Nanobot
+  - Use agent-specific skill paths:
+    - OpenCode: `.agents/skills/` (default)
+    - Claude Code: `.claude/skills/` or `~/.claude/skills/`
+    - OpenClaw: `.openclaw/skills/`
+    - Nanobot: `.nanobot/skills/`
+  - Install skills: saleor-app, saleor-configurator, saleor-core, saleor-storefront
+
+  **References**:
+  - `https://github.com/saleor/agent-skills`
+  - `https://agentskills.io` - Agent Skills Specification
+  - Skill format: SKILL.md + AGENTS.md + rules/ + references/
+
+  **QA Scenarios**:
+  ```
+  Scenario: skills install creates correct directory structure
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js agent skills install
+    Expected: .agents/skills/ directory with saleor-* skills
+    Evidence: .sisyphus/evidence/task-10-skills-structure.{ext}
+
+  Scenario: skills have correct format (SKILL.md, AGENTS.md)
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js agent skills install
+      2. ls .agents/skills/saleor-app/
+    Expected: SKILL.md and AGENTS.md files present
+    Evidence: .sisyphus/evidence/task-10-skills-format.{ext}
+  ```
+
+- [ ] 11. Agent setup - AGENTS.md + IDE detection
+
+  **What to do**:
+  - `jolly agent setup` (combined command)
+  - Detect IDEs/agents present: OpenCode, Claude Code, OpenClaw, Nanobot
+  - Generate AGENTS.md with Saleor conventions (IDE-aware)
+  - Include build/lint/test commands
+  - Create appropriate skill directory structure for each agent
+
+  **Supported Agents**:
+  - OpenCode: `.agents/skills/`
+  - Claude Code: `.claude/skills/` or `~/.claude/skills/`
+  - OpenClaw: `.openclaw/skills/`
+  - Nanobot: `.nanobot/skills/`
+
+  **QA Scenarios**:
+  ```
+  Scenario: agent setup creates AGENTS.md
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js agent setup
+    Expected: AGENTS.md created in current directory
+    Evidence: .sisyphus/evidence/task-11-agents-md.{ext}
+
+  Scenario: AGENTS.md contains Saleor-specific commands
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js agent setup
+      2. grep -i "saleor" AGENTS.md
+    Expected: Saleor commands and conventions present
+    Evidence: .sisyphus/evidence/task-11-agents-content.{ext}
+  ```
+
+- [ ] 12. Agent setup - MCP configuration (IDE-aware)
+
+  **What to do**:
+  - Detect IDE and create appropriate MCP config:
+    - OpenCode: `.mcp.json` in project root
+    - Claude Code: `.mcp.json` in project root (compatible)
+  - Configure saleor-mcp server
+  - Point to https://mcp.saleor.app
+
+  **References**:
+  - `https://mcp.saleor.app`
+  - MCP config format: { "mcpServers": { "saleor": { "command": "...", "url": "..." } } }
+
+  **QA Scenarios**:
+  ```
+  Scenario: agent setup creates .mcp.json with saleor-mcp
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js agent setup
+    Expected: .mcp.json created with saleor-mcp server config
+    Evidence: .sisyphus/evidence/task-12-mcp-config.{ext}
+
+  Scenario: .mcp.json points to correct saleor-mcp URL
+    Tool: interactive_bash
+    Steps:
+      1. ./dist/jolly.js agent setup
+      2. cat .mcp.json | grep mcp.saleor.app
+    Expected: URL https://mcp.saleor.app present
+    Evidence: .sisyphus/evidence/task-12-mcp-url.{ext}
+  ```
+
+- [ ] 13. Integration tests
+
+  **What to do**:
+  - Test store creation flow (mocked API)
+  - Test app scaffold flow
+  - Test agent setup flow
+  - Test error handling
+
+  **QA Scenarios**:
+  ```
+  Scenario: Full integration test suite passes
+    Tool: interactive_bash
+    Steps:
+      1. bun test
+    Expected: All integration tests pass
+    Evidence: .sisyphus/evidence/task-13-integration.{ext}
+
+  Scenario: Error handling tests cover edge cases
+    Tool: interactive_bash
+    Steps:
+      1. bun test src/test/error-handling.test.ts
+    Expected: Error tests pass
+    Evidence: .sisyphus/evidence/task-13-errors.{ext}
+  ```
+
+---
+
+## Final Verification Wave
+
+- [ ] F1. **Plan Compliance Audit** — `oracle`
+- [ ] F2. **Code Quality Review** — `unspecified-high`
+- [ ] F3. **Real Manual QA** — `unspecified-high`
+- [ ] F4. **Scope Fidelity Check** — `deep`
+
+---
+
+## Commit Strategy
+
+- **1**: `init: project scaffolding` — package.json (with bin entries), tsconfig, src/
+- **2**: `feat: CLI structure` — OpenTUI + yargs setup
+- **3**: `feat: Saleor API client` — Cloud API integration
+- **4**: `feat: OpenTUI components` — Rich output components
+- **5**: `feat: bootstrap command` — Store bootstrap command
+- **6**: `feat: agent setup command` — Agent setup command
+- **7**: `feat: app commands` — Dashboard/payment/webhook scaffold
+- **8**: `test: BDD + integration` — Tests
+
+---
+
+## Success Criteria
+
+### Verification Commands
+```bash
+# Build
+bun run build
+
+# Full CLI via npx
+./dist/jolly.js --help
+
+# npm create entry point (store bootstrap)
+./dist/create-saleor-jolly.js --help
+
+# npm init entry point (agent setup)
+./dist/init-saleor-jolly.js --help
+
+# Run tests
+bun test
+```
+
+### Final Checklist
+- [ ] All "Must Have" present
+- [ ] All "Must NOT Have" absent
+- [ ] All tests pass
+- [ ] CLI produces working executable

package/AGENTS.md

@@ -0,0 +1,46 @@
+# Saleor Development Guide
+
+This project uses Saleor e-commerce platform.
+
+## Commands
+
+```bash
+# Development
+npm run dev
+
+# Build
+npm run build
+
+# Test
+npm run test
+
+# Lint
+npm run lint
+```
+
+## Saleor Cloud
+
+- Dashboard: https://cloud.saleor.io
+- Documentation: https://docs.saleor.io
+- API Reference: https://docs.saleor.io/api
+
+## Saleor Skills
+
+This project includes Saleor agent skills:
+- saleor-app: App development patterns
+- saleor-configurator: Config as code
+- saleor-core: Backend internals
+- saleor-storefront: Storefront patterns
+
+## MCP Server
+
+Configure saleor-mcp for AI agent capabilities:
+```json
+{
+  "mcpServers": {
+    "saleor": {
+      "url": "https://mcp.saleor.app"
+    }
+  }
+}
+```