Claude Code cheat sheet

#!/usr/bin/env bash
# Start in plan mode: Claude reads and proposes a plan, no edits until you approve.
# In a session, Shift+Tab cycles modes and /plan <task> enters plan mode.
# Source: https://code.claude.com/docs/en/permission-modes#analyze-before-you-edit-with-plan-mode
claude --permission-mode plan

#!/usr/bin/env bash
# Branch a conversation: resume it under a NEW session id so the original stays untouched.
# Inside a session, /branch does the same thing.
# Source: https://code.claude.com/docs/en/sessions#branch-a-session
claude --continue --fork-session

# Project instructions (CLAUDE.md)

<!-- Block-level HTML comments like this one are stripped before the file is sent to Claude, so they cost no context tokens. -->

## Commands
- Build: `npm run build`
- Test one file: `npm test -- path/to/file.test.ts`
- Lint before every commit: `npm run lint`

## Conventions
- Use 2-space indentation.
- API handlers live in `src/api/handlers/`.
- Use the standard error response format from `src/api/errors.ts`.

## Workflow
- Run `npm test` before committing.
- Never push directly to `main`; open a pull request.

## Compact instructions
When you are using compact, please focus on test output and code changes.

<!-- Keep this file under about 200 lines. Move rules that only matter for some files into .claude/rules/ with `paths:` frontmatter. -->

{
  "$schema": "https://json.schemastore.org/claude-code-settings.json",
  "permissions": {
    "defaultMode": "acceptEdits",
    "allow": [
      "Bash(npm run *)",
      "Bash(git commit *)",
      "Bash(git diff *)",
      "WebFetch(domain:docs.anthropic.com)",
      "mcp__puppeteer__*",
      "Read(~/.zshrc)"
    ],
    "ask": [
      "Bash(git push *)"
    ],
    "deny": [
      "Read(./.env)",
      "Read(./.env.*)",
      "Read(./secrets/**)",
      "Bash(curl *)",
      "Bash(rm -rf *)"
    ],
    "additionalDirectories": [
      "../shared-config"
    ]
  }
}

#!/usr/bin/env bash
# Structured output that must match a JSON Schema. The answer is in .structured_output.
# Check that the field exists before you use it: it is absent if the run ended without valid output.
# Source: https://code.claude.com/docs/en/headless#get-structured-output
claude -p "Extract the main function names from auth.py" \
  --output-format json \
  --json-schema '{"type":"object","properties":{"functions":{"type":"array","items":{"type":"string"}}},"required":["functions"]}' \
  | jq '.structured_output'

#!/usr/bin/env bash
# Parallel, isolated sessions: each name gets its own git worktree and branch (worktree-<name>).
# Needs at least one commit. Add .claude/worktrees/ to .gitignore.
# Source: https://code.claude.com/docs/en/worktrees#start-claude-in-a-worktree
claude --worktree feature-auth          # terminal 1
claude -w bugfix-123                    # terminal 2 (short flag)
claude -w "#1234"                       # branch the worktree from a pull request

#!/usr/bin/env bash
# Adding MCP servers. Every flag below appears in `claude mcp add --help` (v2.1.296)
# and in https://code.claude.com/docs/en/mcp.md#installing-mcp-servers
# Do not run blindly: these add real servers to your config.

# 1) Remote HTTP server (recommended transport)
claude mcp add --transport http notion https://mcp.notion.com/mcp

# 2) Remote HTTP server with a bearer header, saved for ALL your projects
claude mcp add --transport http secure-api https://api.example.com/mcp \
  --scope user \
  --header "Authorization: Bearer your-token"

# 3) Remote SSE server (deprecated transport; HTTP is tried first on v2.1.265+)
claude mcp add --transport sse asana https://mcp.asana.com/sse

# 4) Local stdio server. `--` separates Claude's flags from the server command.
#    Put another option between --env and the server name.
claude mcp add --env AIRTABLE_API_KEY=YOUR_KEY --transport stdio airtable \
  -- npx -y airtable-mcp-server

# 5) Share a server with the team: writes .mcp.json in the project root
claude mcp add --transport http shared-server --scope project https://example.com/mcp

# 6) Pre-registered OAuth: fixed callback port and client id (prompts for the secret)
claude mcp add --transport http --callback-port 8080 --client-id your-client-id \
  --client-secret my-server https://mcp.example.com/mcp

# 7) Paste a JSON block (type is required for url entries: http | sse | ws)
claude mcp add-json weather-api \
  '{"type":"http","url":"https://api.weather.com/mcp","headers":{"Authorization":"Bearer token"}}'

# 8) WebSocket servers can only be added with add-json
claude mcp add-json events-server \
  '{"type":"ws","url":"wss://mcp.example.com/socket","headers":{"Authorization":"Bearer YOUR_TOKEN"}}'

# 9) Import from Claude Desktop (macOS and WSL only)
claude mcp add-from-claude-desktop --scope user

---
name: deploy
description: Deploy the application to staging or production. Use when the user asks to ship, release or deploy a build.
when_to_use: Trigger phrases such as "deploy to staging", "ship it", "release v2".
argument-hint: "[environment] [version]"
arguments: [environment, version]
disable-model-invocation: true
allowed-tools: Bash(${CLAUDE_SKILL_DIR}/scripts/deploy.sh *) Read Grep
model: inherit
effort: high
paths: "src/**,deploy/**"
shell: bash
---

Deploy $environment (version $version) as follows:

1. Run the test suite and stop if anything fails.
2. Run `${CLAUDE_SKILL_DIR}/scripts/deploy.sh $environment $version`.
3. Verify the deployment and report the URL.

Current branch: !`git branch --show-current`

Reference material lives next to this file and loads only when needed:
see [reference.md](reference.md) for the rollback runbook.

Session: ${CLAUDE_SESSION_ID}

---
name: code-reviewer
description: Reviews code for quality, security and best practices. Use proactively after code changes.
tools: Read, Glob, Grep, Bash
disallowedTools: Write, Edit
model: sonnet
permissionMode: default
maxTurns: 25
effort: medium
skills:
  - api-conventions
memory: project
background: false
color: blue
---

You are a senior code reviewer. When invoked:

1. Run `git diff` to see recent changes.
2. Review the modified files for correctness, security and readability.
3. Report findings grouped by severity: critical, warning, suggestion.
4. Update your agent memory with recurring issues you find in this repo.

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Bash",
        "hooks": [
          {
            "type": "command",
            "command": "python3",
            "args": [
              "${CLAUDE_PROJECT_DIR}/.claude/hooks/hook_block_rm.py"
            ],
            "if": "Bash(rm *)",
            "timeout": 10,
            "onFailure": "block"
          }
        ]
      }
    ]
  }
}