[Sizning so'rovingiz]
    ↓
Read / Grep / Bash
    ↓
Think → Edit / Write
    ↓
Observe (natijani ko'radi, kerak bo'lsa qaytadan tuzatadi)

npm install -g @anthropic-ai/claude-code
claude
# Brauzer ochiladi, Anthropic hisobi bilan login. Tamom.

your-project/
├── CLAUDE.md
├── CLAUDE.local.md          # shaxsiy, gitignore'da
├── .mcp.json
└── .claude/
    ├── settings.json        # hooks, statusLine, permissions
    ├── rules/               # jarayon qoidalari
    │   ├── code-style.md
    │   └── testing.md
    ├── commands/            # custom /slash komandalar
    │   └── review.md
    ├── skills/              # qayta ishlatiladigan bilim
    ├── agents/              # maxsus sub-agent'lar
    └── hooks/               # shell script hook'lar

# My Blog Platform

## Stack

- Django 5 + DRF + PostgreSQL 16 + Redis 7
- Celery 5, Docker Compose, MinIO

## Commands

- `make up` — start containers
- `make test` — run pytest
- `make migrate` — run migrations

## Rules

- No N+1 queries — use `select_related`/`prefetch_related`
- Minimum 1 test per view
- Celery tasks only in `tasks.py`

# Testing rules

- Minimum 1 test per DRF view
- Test name: `test_<action>_<expectation>`
- Mock — only for external APIs. DB — real.

---
description: Review the current PR
---

1. Read `git diff main...HEAD`
2. Check for bugs, security issues, SOLID, DRY violations
3. For each issue: file:line, reason, recommendation

/model opusplan

{
  "statusLine": {
    "type": "command",
    "command": "bash ~/.claude/scripts/statusline.sh"
  }
}

#!/usr/bin/env bash
INPUT=$(cat)
MODEL=$(echo "$INPUT" | jq -r '.model.display_name // "claude"')
BRANCH=$(git -C "$(echo "$INPUT" | jq -r '.cwd')" branch --show-current 2>/dev/null || echo "-")
TOKENS=$(echo "$INPUT" | jq -r '.session.total_tokens // 0')
COST=$(echo "$INPUT" | jq -r '.session.total_cost_usd // 0')
printf "📦 %s | 🌿 %s | 🪙 %'d | 💸 \$%.3f" "$MODEL" "$BRANCH" "$TOKENS" "$COST"

{
  "hooks": {
    "PostToolUse": [
      {
        "matcher": "Write|Edit",
        "hooks": [
          {
            "type": "command",
            "command": "ruff format $CLAUDE_TOOL_FILE_PATH 2>/dev/null || true"
          }
        ]
      }
    ]
  }
}

~/.claude/skills/django-test-writer/
├── SKILL.md
└── references/pytest-fixtures.md

---
name: django-test-writer
description: Use when writing or refactoring Django test files.
  Generates pytest-django tests with factory_boy fixtures.
---

1. Read the view file
2. Check existing conftest.py
3. Generate test with auth, permissions, happy-path, edge cases
4. Run pytest -q

---
name: code-reviewer
description: Reviews code for bugs, SOLID violations, security.
tools: Read, Grep, Glob, Bash
---

Bugs, SOLID/DRY violations, security issues.
Return only high-priority problems. No style preferences.

{
  "mcpServers": {
    "context7": {
      "command": "npx",
      "args": ["-y", "@upstash/context7-mcp@latest"]
    },
    "playwright": {
      "command": "npx",
      "args": ["-y", "@playwright/mcp@latest"]
    },
    "postgres": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-postgres",
        "postgres://user:pass@localhost:5432/mydb"
      ]
    }
  }
}

claude mcp add context7 npx @upstash/context7-mcp@latest
claude mcp list

/plugin install superpowers
/plugin install context7
/plugin install playwright

/loop 10m /check-deploy

/loop /fix-flaky-tests

/loop 10m "
docker compose ps ni o'qi.
Agar biror konteyner unhealthy — so'nggi 50 qator log ko'r.
Agar hammasi healthy — 'OK' yozib to'xta.
Maksimum 2 soat.
"

1. cd your-project/
2. claude → /init  (CLAUDE.md avtomatik)
3. CLAUDE.md'ni qisqartiring (<500 qator)
4. .claude/settings.json — statusLine qo'shing
5. .mcp.json — context7, postgres, playwright
6. .claude/rules/ — testing.md, commits.md
7. .claude/hooks/ — ruff format, eslint avtomatik
8. .claude/commands/ — /review, /deploy
9. .claude/skills/ — loyihaga maxsus skill
10. /model opusplan — default qiling