---
name: brand-voice
description: Apply Acme Corp's brand voice to all written content. Use when writing blog posts, marketing copy, social media, or any external-facing content.
---

my-skill/
├── SKILL.md              # Main instructions (required)
├── templates/
│   └── report-template.md
├── examples/
│   └── sample-output.md
└── scripts/
    └── validate.py

mkdir -p ~/.claude/skills/brand-voice
cd ~/.claude/skills/brand-voice

---
name: brand-voice
description: Apply consistent brand voice to all written content. Use when writing blog posts, emails, social media, marketing copy, or any external-facing communication.
---

# Brand Voice Guidelines

When writing any content, apply these voice principles:

## Tone Characteristics

- **Conversational but professional**: Write like you're explaining to a smart colleague
- **Clear over clever**: Avoid jargon unless necessary; explain complex ideas simply
- **Confident but humble**: Share expertise without being condescending
- **Energetic but not hype-driven**: Show enthusiasm through insight, not exclamation points

## Writing Rules

1. **Use active voice** whenever possible
   - Good: "We built this feature"
   - Bad: "This feature was built by us"

2. **Keep paragraphs short** (2-4 sentences max)
   - Break up walls of text
   - Use bullet points for scannable lists

3. **Avoid AI slop phrases**:
   - Don't use: "delve into," "it's worth noting," "importantly," "overall"
   - Don't use: "Not only X, but also Y" constructions
   - Don't overuse em dashes or semicolons

4. **Lead with value**
   - Start with what the reader gains
   - Front-load the most important information

## Examples

**Before:**
"It is important to note that our platform facilitates the optimization of workflows, thereby enabling users to achieve greater efficiency in their daily tasks."

**After:**
"Our platform helps you work faster. It automates repetitive tasks so you can focus on what matters."

## When to Apply

Apply this voice to:
- Blog posts and articles
- Marketing emails
- Product descriptions
- Social media posts
- External presentations
- Customer-facing documentation

Do NOT apply to:
- Technical documentation (stay precise)
- Legal or compliance content (stay formal)
- Code comments (stay concise)

Can you help me write a blog post about our new feature launch?

/brand-voice

Rewrite this paragraph: "Our solution leverages cutting-edge technology to facilitate enhanced productivity outcomes."

mkdir -p ~/.claude/skills/code-review/{templates,examples}
cd ~/.claude/skills/code-review

---
name: code-review
description: Perform thorough code reviews following engineering best practices. Use when reviewing pull requests, code changes, or when asked to analyze code quality.
allowed-tools: Read, Grep
---

# Code Review Skill

Perform comprehensive code reviews focusing on quality, security, and maintainability.

## Review Process

When reviewing code, analyze in this order:

1. **Correctness**: Does the code do what it's supposed to do?
2. **Security**: Are there any vulnerabilities or unsafe patterns?
3. **Performance**: Are there obvious inefficiencies?
4. **Maintainability**: Is the code readable and well-structured?
5. **Testing**: Are there adequate tests?

## Review Template

Use the template in `templates/review-template.md` to structure your feedback.

## Common Issues to Flag

### Security Issues
- Hardcoded credentials or API keys
- SQL injection vulnerabilities
- Unvalidated user input
- Missing authentication/authorization checks

### Code Quality Issues
- Functions longer than 50 lines
- Deeply nested conditionals (>3 levels)
- Duplicate code
- Magic numbers without constants
- Missing error handling

### Performance Issues
- N+1 query problems
- Unnecessary loops or iterations
- Missing database indexes
- Inefficient algorithms

## Output Format

Always structure reviews as:

1. **Summary**: One-sentence verdict (Approve / Request Changes / Comment)
2. **Critical Issues**: Must-fix problems
3. **Suggestions**: Nice-to-have improvements
4. **Positive Feedback**: What was done well

See `examples/sample-review.md` for a complete example.

## When to Use This Skill

- Reviewing pull requests
- Analyzing code quality
- Security audits
- Refactoring suggestions
- Teaching code best practices

# Code Review: [Title]

**Reviewer**: Claude
**Date**: [Current Date]
**Files Reviewed**: [List of files]

## Summary
[One-sentence verdict: Approve / Request Changes / Comment]

## Critical Issues
[Must-fix problems that block merging]

1. **[Issue Title]**
   - **Location**: `file.js:42`
   - **Problem**: [What's wrong]
   - **Fix**: [How to fix it]

## Suggestions
[Nice-to-have improvements]

1. **[Suggestion Title]**
   - **Location**: `file.js:100`
   - **Current**: [Current approach]
   - **Suggested**: [Better approach]
   - **Why**: [Reasoning]

## Positive Feedback
[What was done well]

- [Specific good practice]
- [Well-structured code]
- [Good test coverage]

## Overall Assessment
[Detailed summary of review]

# Code Review: Add User Authentication

**Reviewer**: Claude
**Date**: March 2, 2026
**Files Reviewed**: `auth.js`, `middleware.js`, `user.model.js`

## Summary
Request Changes - Security issues must be addressed before merging.

## Critical Issues

1. **Hardcoded Secret Key**
   - **Location**: `auth.js:15`
   - **Problem**: JWT secret is hardcoded as "mysecret123"
   - **Fix**: Move to environment variable: `process.env.JWT_SECRET`

2. **Missing Input Validation**
   - **Location**: `auth.js:42`
   - **Problem**: Email and password not validated before database query
   - **Fix**: Add validation using a library like Joi or express-validator

## Suggestions

1. **Add Rate Limiting**
   - **Location**: `auth.js:30`
   - **Current**: No rate limiting on login endpoint
   - **Suggested**: Add express-rate-limit middleware
   - **Why**: Prevents brute force attacks

## Positive Feedback

- Excellent test coverage (95%)
- Clear error messages for users
- Well-structured middleware pattern

## Overall Assessment
Strong implementation with good test coverage. The authentication flow is well-designed, but critical security issues must be fixed before this can be merged safely.

/code-review

Please review this file: src/auth/login.js

mkdir -p ~/.claude/skills/research-agent/{scripts,references}
cd ~/.claude/skills/research-agent

---
name: research-agent
description: Conduct comprehensive research on any topic using web search, document analysis, and synthesis. Use when asked to research, investigate, analyze trends, or gather information on a topic.
context: fork
allowed-tools: WebSearch, Read, Scrape
---

# Research Agent

Conduct thorough, well-sourced research on any topic and synthesize findings into actionable insights.

## Research Process

1. **Understand the Question**
   - Clarify the research objective
   - Identify key areas to investigate
   - Determine success criteria

2. **Information Gathering** (Parallel Execution)
   - Search for recent sources (last 12 months preferred)
   - Identify authoritative sources
   - Collect diverse perspectives
   - Scrape relevant content from key URLs

3. **Analysis**
   - Synthesize findings across sources
   - Identify patterns and trends
   - Note contradictions or debates
   - Assess source credibility

4. **Synthesis**
   - Structure findings logically
   - Highlight key insights
   - Provide actionable recommendations
   - Cite all sources

## Output Format

Structure research reports as:

### Executive Summary
[2-3 sentence overview of findings]

### Key Findings
1. **[Finding Title]**
   - Evidence: [Supporting data]
   - Source: [Citation]
   - Implication: [What this means]

### Detailed Analysis
[Comprehensive breakdown by theme/category]

### Contradictions & Debates
[Areas where sources disagree]

### Recommendations
[Actionable next steps based on research]

### Sources
[Full list of citations with URLs and dates]

## Dynamic Context Injection

When researching, automatically inject:
- Current date context: ${CLAUDE_SESSION_ID}
- Search time range: Last 12 months from today
- Source diversity requirements: Minimum 5 distinct sources

## Subagent Execution

This skill runs in a forked context (`context: fork`) which means:
- Isolated from main conversation
- Can spawn parallel research tasks
- Aggregates findings before returning
- Doesn't pollute main context with intermediate steps

## Quality Standards

**Source Requirements:**
- Prefer primary sources over secondary
- Prioritize recent publications (2025-2026)
- Include academic sources when available
- Verify claims across multiple sources

**Analysis Depth:**
- Go 3 layers deep on key topics
- Explore counterarguments
- Quantify findings when possible
- Note confidence levels

## Example Usage

**Input:**
"Research the current state of AI agent frameworks and which ones are gaining traction"

**Output:**
[Comprehensive report following the structure above, with 8-10 sources, trend analysis, and recommendations]

/research-agent

Research the adoption trends of Claude Skills vs ChatGPT Custom GPTs in enterprise settings. Focus on 2025-2026 data.

---
name: deploy-production
description: Deploy application to production environment
disable-model-invocation: true
context: fork
allowed-tools: Shell, Write
---

---
name: brand-voice
description: Apply brand guidelines to all blog posts, marketing copy, and external content. Trigger whenever writing for external audiences.
user-invocable: true
---

---
name: api-conventions
description: Internal API design patterns and conventions for this codebase
user-invocable: false
---

---
name: deep-research
description: Comprehensive research with web search and synthesis
context: fork
agent: explore
allowed-tools: WebSearch, Read, Scrape
---

advanced-skill/
├── SKILL.md                    # Core instructions (300-500 lines)
├── templates/
│   ├── report-template.md      # Output format templates
│   └── email-template.md
├── examples/
│   ├── good-example.md         # Success examples
│   └── bad-example.md          # Anti-patterns
├── references/
│   ├── api-spec.md             # Detailed reference docs
│   └── style-guide.md          # Extended guidelines
└── scripts/
    ├── validate.py             # Executable validation
    └── transform.js            # Data transformation

## API Reference

For detailed endpoint specifications, see `references/api-spec.md`.

Load this file when:
- Implementing API calls
- Debugging API errors
- Designing new endpoints
- Validating request/response formats

name: multi-framework-deploy
description: Deploy applications across AWS, GCP, or Azure

## Deployment Workflow
1. Detect cloud provider from config
2. Load provider-specific reference
3. Execute deployment steps
4. Validate deployment

## Provider Selection
- AWS: See `references/aws-deploy.md`
- GCP: See `references/gcp-deploy.md`
- Azure: See `references/azure-deploy.md`

<!-- Bad: Loads everything always -->
Here are 50 examples of good code reviews...
[10,000 words of examples]

<!-- Good: Load on demand -->
For examples, see `examples/code-reviews.md` (load when user asks for examples or when review is unclear)

context: fork  # Runs in isolated context

my-claude-skills/
├── README.md
├── skills/
│   ├── brand-voice/
│   │   └── SKILL.md
│   ├── code-review/
│   │   ├── SKILL.md
│   │   ├── templates/
│   │   └── examples/
│   └── research-agent/
│       └── SKILL.md
└── docs/
    ├── installation.md
    └── usage-guide.md

# My Claude Skills Collection

A collection of Claude skills for [your use case].

## Skills Included

### Brand Voice
Applies consistent brand guidelines to all writing.
- **Use when**: Writing blog posts, marketing copy, emails
- **Installation**: Copy to `~/.claude/skills/brand-voice/`

### Code Review
Performs security-focused code reviews.
- **Use when**: Reviewing pull requests
- **Installation**: Copy to `~/.claude/skills/code-review/`

## Installation

1. Clone this repository
2. Copy desired skills to your Claude skills directory:
   ```bash
   cp -r skills/brand-voice ~/.claude/skills/

### Documentation Best Practices

**For Each Skill, Document:**

1. **What it does** (one sentence)
2. **When to use it** (specific scenarios)
3. **Example invocations** (show both auto-trigger and manual)
4. **Expected output** (screenshots or examples)
5. **Customization options** (what users might want to change)

**Example Skill Documentation:**

```markdown
## Code Review Skill

**What it does**: Performs comprehensive security-focused code reviews following industry best practices.

**When to use it**:
- Reviewing pull requests before merging
- Security audits of existing code
- Teaching junior developers code quality

**Example invocations**:

Auto-trigger:
"Can you review this PR for security issues?"

Manual:
"/code-review src/auth/login.js"

**Expected output**:
- Summary verdict (Approve/Request Changes)
- Critical security issues with fixes
- Code quality suggestions
- Positive feedback on good practices

**Customization**:
Edit `SKILL.md` to adjust:
- Security rules (lines 25-40)
- Code quality standards (lines 42-58)
- Review template (see templates/review-template.md)