Skip to content

Core Commands: 15 Slash Commands Complete Guide ​

What You'll Learn ​

  • Quickly launch TDD development workflow for high-quality code
  • Create systematic implementation plans to avoid missing key steps
  • Run comprehensive code reviews and security audits
  • Generate end-to-end tests to verify critical user flows
  • Automatically fix build errors and save debugging time
  • Safely clean up dead code and keep your codebase lean
  • Extract and reuse patterns from solved problems
  • Manage work states and checkpoints
  • Run comprehensive verification to ensure code is ready

Your Current Challenges ​

You may encounter these problems during development:

  • Don't know where to start — Facing new requirements, how to break down implementation steps?
  • Low test coverage — Lots of code written but insufficient tests, quality hard to guarantee
  • Build error pile-up — After modifying code, type errors appear one after another, don't know where to fix
  • Non-systematic code review — Manual review easily misses security issues
  • Repeatedly solving the same problems — Falling into the same traps again

Everything Claude Code's 15 slash commands are designed to solve these pain points.

Core Concept ​

Commands are entry points to workflows. Each command encapsulates a complete development process, invoking corresponding agents or skills to help you complete specific tasks.

Commands vs Agents vs Skills

  • Commands: Quick entry points you type directly in Claude Code (like /tdd, /plan)
  • Agents: Specialized sub-agents called by commands, responsible for specific execution
  • Skills: Workflow definitions and domain knowledge that agents can reference

A command typically calls one or more agents, and agents may reference relevant skills.

Commands Overview ​

15 commands categorized by function:

CategoryCommandPurpose
Development Workflow/planCreate implementation plan
/tddExecute test-driven development
/orchestrateExecute multiple agents in sequence
Code Quality/code-reviewCode review
/build-fixFix build errors
/refactor-cleanClean up dead code
/verifyComprehensive verification
Testing/e2eEnd-to-end testing
/test-coverageAnalyze test coverage
/evalManage eval-driven development
Docs & Architecture/update-docsSync documentation
/update-codemapsUpdate architecture docs
State Management/checkpointSave work state
Learning & Optimization/learnExtract reusable patterns
Configuration/setup-pmConfigure package manager

Development Workflow Commands ​

/plan: Plan Complex Features ​

Purpose: Create detailed implementation plans before starting to code.

Use Cases:

  • Starting new feature development
  • Major architectural changes
  • Requirements unclear or ambiguous
  • Multiple files/components coordinated changes

Workflow:

1. Restate requirements
   ↓
2. Identify risks and dependencies
   ↓
3. Break down into specific phases
   ↓
4. Wait for your confirmation

Key Feature

The planner agent will not write any code until you explicitly confirm the plan.

Example:

You: /plan I need to add real-time notification functionality

Claude:
# Implementation Plan: Real-Time Market Resolution Notifications

## Requirements Restatement
- Send notifications when markets resolve
- Support multiple channels (in-app, email, webhook)
- Ensure reliable delivery
- Include market outcome and user's position result

## Implementation Phases
### Phase 1: Database Schema
- Add notifications table
- Add user_notification_preferences table
- Create indexes

### Phase 2: Notification Service
- Create notification queue using BullMQ/Redis
- Implement retry logic
- Create notification templates

...

WAITING FOR CONFIRMATION: Proceed with this plan? (yes/no/modify)

Best Practices:

  • ✅ Plan complex features before implementing
  • ✅ Carefully check risks and dependencies
  • ✅ Execute in phases, verify each phase
  • ❌ Don't skip planning directly

/tdd: Test-Driven Development ​

Purpose: Enforce TDD workflow, ensuring test coverage ≥ 80%.

Use Cases:

  • Implementing new features
  • Fixing bugs (write tests that reproduce the bug first)
  • Refactoring existing code
  • Writing core business logic

TDD Loop:

RED → GREEN → REFACTOR → REPEAT

RED:      Write a failing test
GREEN:    Write minimal code to make test pass
REFACTOR: Improve code while keeping tests passing
REPEAT:   Next feature/scenario

Mandatory Requirement

Must write tests first, then code. Cannot skip RED phase.

Example:

You: /tdd I need a function to calculate market liquidity score

Claude (tdd-guide agent):
## Step 1: Define Interface (SCAFFOLD)
```typescript
export interface MarketData {
  totalVolume: number
  bidAskSpread: number
  activeTraders: number
  lastTradeTime: Date
}

Step 2: Write Failing Test (RED) ​

typescript
test('should return high score for liquid market', () => {
  const market = { ... }
  expect(calculateLiquidityScore(market)).toBeGreaterThan(80)
})

Step 3: Run Tests - Verify FAIL ​

bash
npm test lib/liquidity.test.ts
FAIL ✕ should return high score for liquid market
  Error: Not implemented

Step 4: Implement Minimal Code (GREEN) ​

typescript
export function calculateLiquidityScore(market: MarketData): number {
  // Implementation code...
}

Step 5: Run Tests - Verify PASS ​

bash
npm test lib/liquidity.test.ts
PASS ✓ should return high score for liquid market

Step 6: Refactor (IMPROVE) ​

// Improve code structure...

Step 7: Verify Tests Still Pass ​

bash
npm test lib/liquidity.test.ts
PASS ✓ all tests passing

Step 8: Check Coverage ​

Coverage: 100% ✅ (Target: 80%)


**Best Practices**:
- ✅ Write interface definition first
- ✅ Implement after test fails
- ✅ Test immediately after completing each feature
- ✅ Verify tests pass before and after refactoring
- ✅ Coverage ≥ 80% (100% for critical code)
- ❌ Don't write too much code before testing
- ❌ Don't skip RED phase
- ❌ Don't test implementation details

---

### /orchestrate: Sequential Agent Execution

**Purpose**: Automate sequential workflow of multiple agents.

**Use Cases**:
- Complete feature development (planner → tdd-guide → code-reviewer → security-reviewer)
- Bug fix workflow (explorer → tdd-guide → code-reviewer)
- Security audit (security-reviewer → code-reviewer → architect)

**Workflow Types**:

| Workflow Type | Agent Sequence |
|--- | ---|
| `feature` | planner → tdd-guide → code-reviewer → security-reviewer |
| `bugfix` | explorer → tdd-guide → code-reviewer |
| `refactor` | architect → code-reviewer → tdd-guide |
| `security` | security-reviewer → code-reviewer → architect |

**Example**:

You: /orchestrate feature "Add user authentication functionality"

Claude:

ORCHESTRATION REPORT ​

Workflow: feature Task: 添加用户认证功能 Agents: planner → tdd-guide → code-reviewer → security-reviewer

SUMMARY ​

Successfully implemented user authentication functionality, including registration, login, session management.

AGENT OUTPUTS ​

Planner: Created 4-phase implementation plan TDD Guide: Implemented auth service, 95% coverage Code Reviewer: Found 3 high-priority issues, fixed Security Reviewer: Passed security audit, no vulnerabilities

FILES CHANGED ​

  • src/services/auth.ts
  • src/api/auth/route.ts
  • tests/auth.test.ts
  • ...

TEST RESULTS ​

PASS: 24/24 tests (100%)

SECURITY STATUS ​

✅ No vulnerabilities found

RECOMMENDATION ​

✅ SHIP - Ready for production


**Best Practices**:
- ✅ Use orchestrate for complex features
- ✅ Clear handoff documents between agents
- ✅ Parallel execution for independent checks (code-reviewer and security-reviewer)
- ❌ Don't use orchestrate for simple tasks

---

## Code Quality Commands

### /code-review: Code Review

**Purpose**: Comprehensively review uncommitted code, checking for security issues and code quality.

**Use Cases**:
- Before committing code
- Before merging PR
- After completing feature development

**Check Items**:

| Category | Check Content | Severity |
|--- | --- | ---|
| **Security** | Hardcoded keys, SQL injection, XSS, missing input validation | CRITICAL |
| **Code Quality** | Functions > 50 lines, files > 800 lines, nesting depth > 4 | HIGH |
| **Best Practices** | Mutation patterns, console.log, missing tests, TODO/FIXME | MEDIUM |

**Example**:

You: /code-review

Claude (code-reviewer agent):

Code Review Report ​

Security Issues (CRITICAL) ​

🔴 [HIGH] Hardcoded API key File: src/api/github.ts:15 Issue: GitHub token is hardcoded Fix: Move to environment variable

Code Quality (HIGH) ​

🟡 [MEDIUM] Function exceeds 50 lines File: src/utils/formatter.ts:62 Lines: 78 Issue: FormatLargeData function is too long Fix: Split into smaller functions

Best Practices (MEDIUM) ​

🟢 [LOW] TODO comment File: src/services/user.ts:45 Issue: TODO: Add rate limiting Fix: Implement or remove comment

SUMMARY ​

CRITICAL: 1 (Must fix before commit) HIGH: 2 (Should fix) MEDIUM: 3 (Nice to have)

❌ BLOCKED - Critical security issues found


::: danger Block Commit Rule

If **CRITICAL** or **HIGH** level issues are found, commits will be blocked.

:::

---

### /build-fix: Fix Build Errors

**Purpose**: Fix TypeScript and build errors one by one.

**Use Cases**:
- `npm run build` fails
- TypeScript type errors
- Compilation errors

**Workflow**:
  1. Run build ↓
  2. Parse error output, sort by severity ↓
  3. Fix one by one:
    • Show error context
    • Explain problem
    • Propose fix
    • Apply fix
    • Re-run build ↓
  4. Generate fix summary

**Example**:

You: /build-fix

Claude (build-error-resolver agent):

Fixing Build Errors ​

Running build: npm run build Found 5 errors

Error 1: Type 'string' is not assignable to type 'number' ​

File: src/calculator.ts:15

typescript
const result = add("10", "20")  // Error

Issue: add function expects numbers, received strings Fix: Convert to numbers

typescript
const result = add(Number("10"), Number("20"))

✅ Fixed. Re-running build...

Error 2: Cannot find module './utils' ​

File: src/index.ts:3 Issue: Import path is incorrect Fix: Change to './utils/helper'

typescript
import { helper } from './utils/helper'

✅ Fixed. Re-running build...

...

SUMMARY ​

Errors Fixed: 5 Errors Remaining: 0 Build Status: ✅ PASS


::: tip Safety First

Fix only one error at a time to prevent introducing new issues.

:::

---

### /refactor-clean: Clean Up Dead Code

**Purpose**: Safely identify and remove unused code.

**Use Cases**:
- Large amounts of unused files and exports in codebase
- Reduce maintenance burden
- Improve code readability

**Workflow**:
  1. Run dead code analysis tools
    • knip: Find unused exports and files
    • depcheck: Find unused dependencies
    • ts-prune: Find unused TypeScript exports ↓
  2. Generate comprehensive report ↓
  3. Classify by severity
    • SAFE: Test files, unused utility functions
    • CAUTION: API routes, components
    • DANGER: Config files, entry files ↓
  4. Delete one by one (test first) ↓
  5. Generate cleanup summary

**Example**:

You: /refactor-clean

Claude (refactor-cleaner agent):

Dead Code Analysis ​

Tools Run ​

  • ✅ knip: Found 12 unused exports
  • ✅ depcheck: Found 5 unused dependencies
  • ✅ ts-prune: Found 8 unused TypeScript exports

Classification ​

SAFE (Can delete) ​

  • src/utils/old-helper.ts - Unused utility function
  • src/components/Button.tsx - Not used anywhere
  • lodash - Unused dependency

CAUTION (Review first) ​

  • src/api/users.ts - Only used in tests
  • @types/node - May be used in build scripts

DANGER (Do not delete) ​

  • src/index.ts - Main entry point
  • tsconfig.json - Configuration file

Deletion Plan ​

Run full test suite first: ✅ Tests passing (98% coverage)

Deleting SAFE items...

  1. Removing src/utils/old-helper.ts ✅ Tests still passing

  2. Removing src/components/Button.tsx ✅ Tests still passing

  3. Removing lodash from package.json ✅ Tests still passing

SUMMARY ​

Files deleted: 12 Dependencies removed: 5 Tests still passing: ✅


::: warning Test First

**Must run tests before deleting code** to ensure existing functionality isn't broken.

:::

---

### /verify: Comprehensive Verification

**Purpose**: Run complete code verification checks.

**Use Cases**:
- Before committing code
- Before merging PR
- When preparing for release

**Check Items** (in order):

1. **Build check** - Run build command
2. **Type check** - Run TypeScript checker
3. **Lint check** - Run linter
4. **Test suite** - Run all tests, report coverage
5. **Console.log audit** - Search source files for console.log
6. **Git status** - Show uncommitted changes

**Example**:

You: /verify pre-pr

Claude:

Verification Report ​

VERIFICATION: ✅ PASS

Build: ✅ OK Types: ✅ OK (0 errors) Lint: ✅ OK (0 issues) Tests: ✅ 142/145 passed, 98% coverage Secrets: ✅ OK (0 found) Logs: ⚠️ 3 console.logs found

Console.log locations:

  • src/utils/logger.ts:15 (intentional logger)
  • src/api/webhook.ts:42 (debug, remove before commit)
  • src/services/auth.ts:89 (debug, remove before commit)

Ready for PR: ⚠️ YES (remove debug logs)

Files modified since last commit:

  • src/api/webhook.ts (new)
  • src/services/auth.ts (modified)
  • tests/auth.test.ts (modified)

**Parameters**:
- `quick` - Only build + type check
- `full` - All checks (default)
- `pre-commit` - Commit-related checks
- `pre-pr` - Full checks + security scan

---

## Testing Commands

### /e2e: End-to-End Testing

**Purpose**: Generate and run end-to-end tests using Playwright.

**Use Cases**:
- Test critical user flows (login, trading, payment)
- Verify end-to-end integration of multi-step flows
- Test UI interactions and navigation
- Prepare for production deployment

**Workflow**:
  1. Analyze user flows, identify test scenarios ↓
  2. Generate Playwright tests (using Page Object pattern) ↓
  3. Run tests across multiple browsers (Chrome, Firefox, Safari) ↓
  4. Capture screenshots, videos, traces on failure ↓
  5. Generate report with results and artifacts ↓
  6. Identify flaky tests and suggest fixes

**Example**:

You: /e2e Test market search and view flow

Claude (e2e-runner agent):

E2E Test Generation: Market Search and View Flow ​

Test Scenario Identified ​

User Journey: Market Search → View Results → Click Market → View Details

Generated Test Code ​

typescript
import { test, expect } from '@playwright/test'
import { MarketsPage } from '../../pages/MarketsPage'

test('user can search markets and view details', async ({ page }) => {
  const marketsPage = new MarketsPage(page)
  await marketsPage.goto()

  // Perform search
  await marketsPage.searchMarkets('election')
  await page.waitForResponse(resp =>
    resp.url().includes('/api/markets/search') && resp.status() === 200
  )

  // Verify results
  const marketCards = marketsPage.marketCards
  await expect(marketCards.first()).toBeVisible()

  // Click on first result
  await marketCards.first().click()

  // Verify market details page loads
  await expect(page).toHaveURL(/\/markets\/[a-z0-9-]+/)
  await expect(page.locator('[data-testid="market-name"]')).toBeVisible()
})

Running Tests ​

bash
npx playwright test tests/e2e/markets/search-and-view.spec.ts

✓ [chromium] user can search markets and view details (4.2s)
✓ [chromium] search with no results shows empty state (1.8s)

2 passed (6.0s)

Artifacts generated:
- artifacts/search-results.png
- artifacts/market-details.png
- playwright-report/index.html

Test Report ​

╔══════════════════════════════════════════════════════╗ ║ Status: ✅ ALL TESTS PASSED ║ ║ Total: 2 tests ║ ║ Passed: 2 (100%) ║ ║ Flaky: 0 ║ ╚══════════════════════════════════════════════════════╝


**Quick Commands**:

```bash
# Run all E2E tests
npx playwright test

# Run specific test file
npx playwright test tests/e2e/markets/search.spec.ts

# Run in headed mode (see browser)
npx playwright test --headed

# Debug test
npx playwright test --debug

# View report
npx playwright show-report

Best Practices:

  • ✅ Use Page Object Model for maintainability
  • ✅ Use data-testid attributes as selectors
  • ✅ Wait for API responses, not arbitrary timeouts
  • ✅ Test critical user flows
  • ❌ Don't use fragile selectors (CSS classes change)
  • ❌ Don't test implementation details

/test-coverage: Test Coverage Analysis ​

Purpose: Analyze test coverage and generate missing tests.

Use Cases:

  • Coverage below 80%
  • Need to supplement tests

Workflow:

1. Run tests with coverage
   ↓
2. Analyze coverage report (coverage/coverage-summary.json)
   ↓
3. Identify files below 80% coverage threshold
   ↓
4. For each under-covered file:
   - Analyze untested code paths
   - Generate unit tests for functions
   - Generate integration tests for APIs
   ↓
5. Verify new tests pass
   ↓
6. Show before/after coverage metrics

Example:

You: /test-coverage

Claude:
# Test Coverage Analysis

Running tests with coverage...
npm test --coverage

## Coverage Report

File                      | % Stmts | % Branch | % Funcs | % Lines
--------------------------|---------|----------|---------|--------
src/services/auth.ts       |   92    |   88     |   100   |   91
src/utils/formatter.ts     |   65    |   50     |   75    |   63  ⚠️
src/api/users.ts          |   100   |   100    |   100   |   100
src/components/Button.tsx  |   45    |   30     |   50    |   44  ⚠️

Overall                   |   78    |   72     |   85    |   76  ⚠️

Target: 80%
Current: 76% ❌ Need to improve

## Generating Missing Tests

### File: src/utils/formatter.ts (65% coverage)
Untested paths:
- FormatCurrency with negative values
- FormatDate with custom locale
- TruncateText with long strings

Generating tests...
```typescript
describe('formatter', () => {
  it('should format negative currency', () => {
    expect(formatCurrency(-100)).toBe('-$100.00')
  })

  it('should format date with custom locale', () => {
    expect(formatDate(new Date(), 'ja-JP')).toMatch('2024年')
  })

  it('should truncate long text', () => {
    expect(truncateText('a'.repeat(100), 10)).toBe('aaaaaaaaaa...')
  })
})

File: src/components/Button.tsx (45% coverage) ​

Untested paths:

  • Disabled state
  • Loading state
  • onClick handler

Generating tests...

After Adding Tests ​

Running tests again...

File% Stmts% Branch% Funcs% Lines
src/services/auth.ts928810091
src/utils/formatter.ts1009510099 ✅
src/api/users.ts100100100100
src/components/Button.tsx959010094 ✅

Overall | 95 | 92 | 100 | 94 ✅

Target: 80% Current: 94% ✅ EXCEEDED

Summary: +12 tests added, Coverage: 76% → 94%


---

### /eval: Eval-Driven Development

**Purpose**: Manage eval-driven development workflow.

**Use Cases**:
- Eval-driven development (EDD)
- Feature acceptance testing
- Regression testing

**Subcommands**:

- `define <name>` - Create new eval definition
- `check <name>` - Run and check evals
- `report <name>` - Generate comprehensive report
- `list` - Show all evals

**Example - Define Eval**:

You: /eval define user-auth

Claude: Creating eval definition...

.claude/evals/user-auth.md:

EVAL: user-auth ​

Created: 2026-01-25

Capability Evals ​

  • [ ] User can register with email and password
  • [ ] User can login with correct credentials
  • [ ] User cannot login with wrong password
  • [ ] User can reset password via email

Regression Evals ​

  • [ ] Existing users can still login
  • [ ] Session tokens still work
  • [ ] OAuth login still works

Success Criteria ​

  • pass@3 > 90% for capability evals
  • pass^3 = 100% for regression evals

Please fill in specific evaluation criteria...


**Example - Check Eval**:

You: /eval check user-auth

Claude: Running evals...

EVAL CHECK: user-auth ​

Capability: 2/4 passing Regression: 2/3 passing Status: IN PROGRESS

Capability Results: ✅ User can register with email and password ✅ User can login with correct credentials ❌ User cannot login with wrong password (fails intermittently) ❌ User can reset password via email (not implemented)

Regression Results: ✅ Existing users can still login ✅ Session tokens still work ❌ OAuth login still works (test failing due to rate limit)

RECOMMENDATION: NEEDS WORK - Fix password validation and implement password reset


**Example - Report Eval**:

You: /eval report user-auth

Claude: Generating comprehensive report...

EVAL REPORT: user-auth ​

Generated: 2026-01-25

CAPABILITY EVALS ​

[login-credentials-pass]: PASS (pass@2) - required retry [login-wrong-pass-fail]: FAIL - see notes [password-reset-pass]: FAIL - not implemented

REGRESSION EVALS ​

METRICS ​

Capability pass@1: 50% Capability pass@3: 100% Regression pass^3: 100%

NOTES ​

Password validation test is flaky due to timing issue. Need to implement password reset flow.

RECOMMENDATION ​

NEEDS WORK - Complete password reset implementation


---

## Documentation & Architecture Commands

### /update-docs: Sync Documentation

**Purpose**: Sync documentation from source code.

**Use Cases**:
- Update documentation after code changes
- Keep documentation in sync with code

**Workflow**:
  1. Read package.json scripts section
    • Generate script reference table
    • Extract descriptions from comments ↓
  2. Read .env.example
    • Extract all environment variables
    • Record purpose and format ↓
  3. Generate docs/CONTRIB.md
    • Development workflow
    • Available scripts
    • Environment setup
    • Testing procedures ↓
  4. Generate docs/RUNBOOK.md
    • Deployment procedures
    • Monitoring and alerts
    • Common issues and fixes
    • Rollback procedures ↓
  5. Identify outdated documentation
    • Find docs not modified in 90+ days
    • List for manual review ↓
  6. Show diff summary

::: tip Single Source of Truth

package.json and .env.example are the single source of truth.

:::

---

### /update-codemaps: Update Architecture Documentation

**Purpose**: Analyze codebase structure and update architecture documentation.

**Use Cases**:
- After code structure changes
- Need to update architecture docs

**Workflow**:
  1. Scan all source files for imports, exports, and dependencies ↓
  2. Generate concise codemaps:
    • codemaps/architecture.md - Overall architecture
    • codemaps/backend.md - Backend structure
    • codemaps/frontend.md - Frontend structure
    • codemaps/data.md - Data models and schemas ↓
  3. Calculate diff percentage from previous version ↓
  4. If change > 30%, request user approval before updating ↓
  5. Add freshness timestamp to each codemap ↓
  6. Save report to .reports/codemap-diff.txt

**Example**:

You: /update-codemaps

Claude: Analyzing codebase structure...

Scanning 156 files... Found 89 imports and 124 exports

Codemap Diff ​

Previous version: 2026-01-20 Current version: 2026-01-25 Diff: 12% (under 30% threshold)

✅ Safe to update

Updated Codemaps ​

codemaps/architecture.md

  • Updated module dependencies
  • Added new services: auth, notifications
  • Removed deprecated modules: legacy-db

codemaps/backend.md

  • Updated API routes structure
  • Added middleware documentation

codemaps/frontend.md

  • Updated component hierarchy
  • Added new pages: profile, settings

codemaps/data.md

  • Updated database schemas
  • Added new tables: users, sessions

Summary: 4 codemaps updated, 12% change


---

## State Management Commands

### /checkpoint: Save Work State

**Purpose**: Create or verify checkpoints in workflows.

**Use Cases**:
- Save important development stages
- Support rollback and progress tracking

**Subcommands**:

- `create <name>` - Create named checkpoint
- `verify <name>` - Verify checkpoint
- `list` - Show all checkpoints
- `clear` - Remove old checkpoints (keep last 5)

**Example**:

You: /checkpoint create feature-start

Claude:

Creating Checkpoint: feature-start ​

Running /verify quick...

Build: ✅ OK Types: ✅ OK

Creating checkpoint...

✅ Checkpoint "feature-start" created ├─ Timestamp: 2026-01-25-14:30 ├─ Git SHA: abc1234 └─ Status: Clean (no uncommitted changes)

You: /checkpoint verify feature-start

Claude: CHECKPOINT COMPARISON: feature-start ​

Files changed: 5 Tests: +3 passed / -0 failed Coverage: +5% / -0% Build: PASS

Changes since checkpoint: ✅ src/services/auth.ts (modified) ✅ tests/auth.test.ts (added) ✅ src/components/Login.tsx (added)


**Workflow**:

[Start] → /checkpoint create "feature-start" | [Implement] → /checkpoint create "core-done" | [Test] → /checkpoint verify "core-done" | [Refactor] → /checkpoint create "refactor-done" | [PR] → /checkpoint verify "feature-start"


---

## Learning & Optimization Commands

### /learn: Extract Reusable Patterns

**Purpose**: Analyze current session and extract patterns that can be saved as skills.

**Use Cases**:
- After solving non-trivial problems
- Discover reusable error resolution patterns
- Record project-specific conventions

**Extraction Content**:

1. **Error Resolution Patterns**
   - What error occurred?
   - What was the root cause?
   - How was it fixed?
   - Is it reusable for similar errors?

2. **Debugging Techniques**
   - Non-obvious debugging steps
   - Effective tool combinations
   - Diagnostic patterns

3. **Workarounds**
   - Library quirks
   - API limitations
   - Version-specific fixes

4. **Project-Specific Patterns**
   - Discovered codebase conventions
   - Architectural decisions made
   - Integration patterns

**Example**:

You: /learn

Claude:

Extracting Patterns from Session ​

Reviewing session history...

Found 3 extractable patterns:

Pattern 1: Supabase Auth Session Recovery ​

Type: Error Resolution Reusability: High (affects all auth endpoints)

Problem ​

After user logs in, subsequent API calls fail with 401 Unauthorized.

Root Cause ​

Supabase session cookie is not being sent to backend API due to CORS configuration.

Solution ​

typescript
// Configure Supabase with correct URL
const supabase = createClient(
  import.meta.env.VITE_SUPABASE_URL,
  import.meta.env.VITE_SUPABASE_ANON_KEY,
  {
    auth: {
      storageKey: 'my-app-auth',
      autoRefreshToken: true,
      detectSessionInUrl: true,
    },
  }
)

When to Use ​

Trigger: Auth endpoints returning 401 unexpectedly


Pattern 2: Playwright Test Flakiness Fix ​

Type: Debugging Technique Reusability: Medium (affects async UI tests)

Problem ​

Test "user can click submit button" fails intermittently with timeout.

Root Cause ​

Button is disabled initially due to async validation, but test doesn't wait.

Solution ​

typescript
// Use explicit wait with predicate
await page.waitForSelector('[data-testid="submit-btn"]:not(:disabled)', {
  timeout: 10000
})

// Then click
await page.locator('[data-testid="submit-btn"]').click()

When to Use ​

Trigger: Test timing issues with async UI state


Do you want to save these patterns? (yes/no/modify)


::: tip Focus Principle

Extract one pattern at a time to keep skills focused.

:::

---

## Configuration Commands

### /setup-pm: Configure Package Manager

**Purpose**: Configure preferred package manager for project or global use.

**Use Cases**:
- Detect current package manager
- Set global or project preferences

**Detection Priority**:

1. Environment variable: `CLAUDE_PACKAGE_MANAGER`
2. Project config: `.claude/package-manager.json`
3. package.json: `packageManager` field
4. Lock files: package-lock.json, yarn.lock, pnpm-lock.yaml, bun.lockb
5. Global config: `~/.claude/package-manager.json`
6. Fallback: First available package manager

**Supported Package Manager Priority**: pnpm > bun > yarn > npm

**Example**:

```bash
# Detect current package manager
node scripts/setup-package-manager.js --detect

# Set global preference
node scripts/setup-package-manager.js --global pnpm

# Set project preference
node scripts/setup-package-manager.js --project bun

# List available package managers
node scripts/setup-package-manager.js --list

Configuration Files:

Global config (~/.claude/package-manager.json):

json
{
  "packageManager": "pnpm"
}

Project config (.claude/package-manager.json):

json
{
  "packageManager": "bun"
}

Environment variable overrides all detection methods:

bash
# macOS/Linux
export CLAUDE_PACKAGE_MANAGER=pnpm

# Windows (PowerShell)
$env:CLAUDE_PACKAGE_MANAGER = "pnpm"

Command Combination Workflows ​

Complete Feature Development Workflow ​

1. /plan "Add user authentication functionality"
   ↓ Create implementation plan
2. /tdd "Implement auth service"
   ↓ TDD development
3. /test-coverage
   ↓ Ensure coverage ≥ 80%
4. /code-review
   ↓ Code review
5. /verify pre-pr
   ↓ Comprehensive verification
6. /checkpoint create "auth-feature-done"
   ↓ Save checkpoint
7. /update-docs
   ↓ Update documentation
8. /update-codemaps
   ↓ Update architecture docs

Bug Fix Workflow ​

1. /checkpoint create "bug-start"
   ↓ Save current state
2. /orchestrate bugfix "Fix login error"
   ↓ Automated bug fix workflow
3. /test-coverage
   ↓ Ensure test coverage
4. /verify quick
   ↓ Verify fix
5. /checkpoint verify "bug-start"
   ↓ Compare with checkpoint

Security Audit Workflow ​

1. /orchestrate security "Audit payment flow"
   ↓ Security-first review workflow
2. /e2e "Test payment flow"
   ↓ End-to-end testing
3. /code-review
   ↓ Code review
4. /verify pre-pr
   ↓ Comprehensive verification

Command Comparison Cheat Sheet ​

CommandMain PurposeTriggers AgentOutput
/planCreate implementation planplannerPhased plan
/tddTDD developmenttdd-guideTests + implementation + coverage
/orchestrateExecute agents in sequenceMultiple agentsComprehensive report
/code-reviewCode reviewcode-reviewer, security-reviewerSecurity and quality report
/build-fixFix build errorsbuild-error-resolverFix summary
/refactor-cleanClean up dead coderefactor-cleanerCleanup summary
/verifyComprehensive verificationBashVerification report
/e2eEnd-to-end testinge2e-runnerPlaywright tests + artifacts
/test-coverageAnalyze coverageBashCoverage report + missing tests
/evalEval-driven developmentBashEval status report
/checkpointSave stateBash + GitCheckpoint report
/learnExtract patternscontinuous-learning skillSkill file
/update-docsSync documentationdoc-updater agentDocumentation update
/update-codemapsUpdate architecturedoc-updater agentCodemap update
/setup-pmConfigure package managerNode.js scriptPackage manager detection

Common Pitfalls ​

❌ Don't Skip Planning Phase ​

For complex features, starting to code directly leads to:

  • Missing important dependencies
  • Architectural inconsistencies
  • Misunderstanding of requirements

✅ Correct Approach: Use /plan to create a detailed plan, wait for confirmation before implementing.


❌ Don't Skip RED Phase in TDD ​

Writing code first then tests is not TDD.

✅ Correct Approach: Strictly follow RED → GREEN → REFACTOR loop.


❌ Don't Ignore CRITICAL Issues from /code-review ​

Security vulnerabilities can lead to serious consequences like data breaches and financial loss.

✅ Correct Approach: Fix all CRITICAL and HIGH level issues before committing.


❌ Don't Delete Code Without Testing ​

Dead code analysis may have false positives, direct deletion may break functionality.

✅ Correct Approach: Run tests before each deletion to ensure existing functionality isn't broken.


❌ Don't Forget to Use /learn ​

Not extracting patterns after solving non-trivial problems means solving the same problem from scratch next time.

✅ Correct Approach: Regularly use /learn to extract reusable patterns and accumulate knowledge.


Summary ​

Everything Claude Code's 15 slash commands provide complete development workflow support:

  • Development Workflow: /plan → /tdd → /orchestrate
  • Code Quality: /code-review → /build-fix → /refactor-clean → /verify
  • Testing: /e2e → /test-coverage → /eval
  • Documentation & Architecture: /update-docs → /update-codemaps
  • State Management: /checkpoint
  • Learning & Optimization: /learn
  • Configuration: /setup-pm

Master these commands to complete development work efficiently, safely, and with quality.


Coming Up Next ​

In next lesson, we'll learn Core Agents Overview.

You'll learn:

  • Responsibilities and use cases for 9 specialized agents
  • When to call which agent
  • How agents collaborate
  • How to customize agent configurations

Appendix: Source Code Reference ​

Click to expand source code locations

Last updated: 2026-01-25

FeatureFile PathLines
TDD Commandcommands/tdd.md1-327
Plan Commandcommands/plan.md1-114
Code Review Commandcommands/code-review.md1-41
E2E Commandcommands/e2e.md1-364
Build Fix Commandcommands/build-fix.md1-30
Refactor Clean Commandcommands/refactor-clean.md1-29
Learn Commandcommands/learn.md1-71
Checkpoint Commandcommands/checkpoint.md1-75
Verify Commandcommands/verify.md1-60
Test Coverage Commandcommands/test-coverage.md1-28
Setup PM Commandcommands/setup-pm.md1-81
Update Docs Commandcommands/update-docs.md1-32
Orchestrate Commandcommands/orchestrate.md1-173
Update Codemaps Commandcommands/update-codemaps.md1-18
Eval Commandcommands/eval.md1-121
Plugin Definition.claude-plugin/plugin.json1-28

Key Constants:

  • TDD coverage target: 80% (100% for critical code) - commands/tdd.md:293-300

Key Functions:

  • TDD loop: RED → GREEN → REFACTOR - commands/tdd.md:40-47
  • Plan wait confirmation mechanism - commands/plan.md:96
  • Code Review severity levels: CRITICAL, HIGH, MEDIUM - commands/code-review.md:33