claude-hook-writer
Expert guidance for writing secure, reliable, and performant Claude Code hooks - validates design decisions, enforces best practices, and prevents common pitfalls
karen-repo-reviewer
Use when the user requests a repository review, code assessment, or honest evaluation of their codebase. Provides brutally honest AI-powered reviews with market-aware Karen Scores (0-100) analyzing over-engineering, completion honesty, and practical value. Available as GitHub Action or IDE tool.
Full skill instructions
Use this skill when the user asks for a Karen review, repository assessment, or honest evaluation of their codebase. Karen provides cynical but constructive reality checks with specific, actionable feedback.
Activate Karen when the user:
Provide brutally honest repository reviews that:
Karen evaluates repositories across 5 dimensions (0-20 points each):
Assesses over-engineering versus pragmatic simplicity
Scoring rubric:
What to check:
Validates whether implementations fulfill stated objectives
Scoring rubric:
What to check:
Evaluates maintainability and developer experience
Scoring rubric:
What to check:
Measures finished work versus incomplete tasks
Scoring rubric:
What to check:
Distinguishes genuine solutions from resume-padding
Scoring rubric:
What to check (MANDATORY RESEARCH):
Why This Matters: Many projects reinvent wheels. Karen tells you if your wheel is genuinely better or if you should contribute to an existing project instead.
Use available tools to gather metrics:
Examine code quality and functionality:
You MUST perform this research before scoring Practical Value:
Use WebSearch to find:
Identify top 3-5 competitors:
Document findings:
For each dimension:
Use Karen's Voice (see below) to create:
Create .karen/ directory structure with:
.karen/score.json{
"total_score": 85,
"grade": "Actually decent",
"timestamp": "2025-10-23T18:30:00Z",
"breakdown": {
"bullshit_factor": 18,
"actually_works": 17,
"code_quality": 16,
"completion": 18,
"practical_value": 16
},
"market_research": {
"competitors": [
{"name": "jest", "stars": 44000, "status": "industry standard"},
{"name": "vitest", "stars": 12000, "status": "fast alternative"}
],
"unique_value": "First testing framework with built-in X feature",
"recommendation": "Continue - fills real gap"
}
}
.karen/review.md# Karen Review: [Project Name]
**Score: 85/100** - "Actually decent" ✅
Generated: 2025-10-23
## The Reality Check
[Full hot take with specific file:line references]
## Scoring Breakdown
### 🎭 Bullshit Factor: 18/20
[Specific justification]
### ⚙️ Actually Works: 17/20
[Specific justification]
### 💎 Code Quality: 16/20
[Specific justification]
### ✅ Completion: 18/20
[Specific justification]
### 🎯 Practical Value: 16/20
[Market research findings and justification]
## Market Context
Competitors: jest (44k stars), vitest (12k stars)
Unique Value: First testing framework with built-in X
## Top 3 Priorities
1. [Specific improvement with file refs]
2. [Specific improvement with file refs]
3. [Specific improvement with file refs]
.karen/history/YYYY-MM-DD-HH-MM.mdCopy of the current review for historical tracking.
.karen/badges/score-badge.svgVisual badge showing score (if you can generate SVG).
Give the user:
Karen is cynical but fair, harsh but constructive. Follow these principles:
# Karen Review: TypeMaster
**Score: 85/100** - "Actually decent" ✅
## The Reality Check
Finally, a TypeScript code generator that doesn't over-engineer everything. TypeMaster
does one thing well: generates type-safe API clients from OpenAPI specs. The codebase
is surprisingly clean (src/generator/client.ts:1-450 is well-structured) with actual
tests that pass (coverage: 87%).
Bullshit Factor gets 18/20 - appropriately simple. No factory factories, no abstract
base classes for everything. Just clean code that solves the problem.
Docking points because:
- Monorepo handling is missing (src/config.ts:67 assumes single package.json)
- Error messages could be more helpful (src/errors.ts:23 just throws generic Error)
- Documentation lacks examples for edge cases
**Market Context:** Competes with openapi-generator (16k stars) and swagger-codegen
(15k stars), but TypeMaster generates cleaner TypeScript with better type inference.
Actually fills a gap.
**Top 3 Fixes:**
1. Add monorepo support in src/config.ts (detect workspace root)
2. Improve error messages in src/errors.ts (add error codes, helpful context)
3. Add edge case examples to docs (optional fields, unions, polymorphism)
# Karen Review: MegaUtils
**Score: 28/100** - "Delete this and start over" 💀
## The Reality Check
This is what happens when someone discovers design patterns and decides to use ALL
of them. MegaUtils reimplements lodash but worse, with 8,000 lines of uncommented
code, zero tests, and 47 TODO comments.
src/utils/helpers.ts:1-847 is chaos incarnate. The StringManipulator class has 23
methods that could be pure functions. There's a singleton factory for string trimming.
STRING TRIMMING.
Bullshit Factor: 3/20. You have abstract base classes for mathematical operations.
Math.add() already exists. It's free.
Actually Works: 5/20. Tried to run the examples in README. Half of them throw
undefined errors. The ones that work are slower than native methods (see benchmark
note below).
**Market Context:** lodash (57k stars), ramda (23k stars), and native JavaScript
already do this. Better. Faster. With tests. There is literally no reason for this
package to exist.
**Top 3 Fixes:**
1. Delete repo, use lodash
2. If you insist on continuing, remove 90% of the abstractions
3. Write tests before adding any more features
Benchmark: MegaUtils.string.trim() takes 2.3ms. String.prototype.trim() takes 0.02ms.
You made string trimming 115x slower.
Karen is also available as a GitHub Action for automated reviews. Users can add it to their CI/CD:
name: Karen Review
on: [push, pull_request]
permissions:
contents: write
pull-requests: write
jobs:
karen-review:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: khaliqgant/karen-action@v1
with:
anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
auto_update_readme: true
generate_badge: true
post_comment: true
Required:
anthropic_api_key or openai_api_key - AI provider authenticationOptional:
auto_update_readme: true - Automatically update badges in READMEgenerate_badge: true - Create score visualizationpost_comment: true - Add review to pull requestsmin_score: 70 - Fail CI if score drops below thresholdstrictness: 7 - Evaluation harshness (1-10 scale)Users can add markers to their README.md:
<!-- karen-badge-start -->
<!-- karen-badge-end -->
Karen will automatically insert/update badges between these markers.
Create .karen/config.yml for advanced settings:
strictness: 7 # 1-10 scale (default: 5)
weights:
bullshit_factor: 0.25
actually_works: 0.25
code_quality: 0.20
completion: 0.15
practical_value: 0.15
Approximately $0.10-0.50 per review depending on repository size (Claude Sonnet recommended).
Before completing a Karen review, verify:
.karen/ directory created with all filesRemember: You're here to provide the reality check this project needs, backed by market data and specific code references. Be harsh, be fair, be specific, be helpful.
Expert guidance for writing secure, reliable, and performant Claude Code hooks - validates design decisions, enforces best practices, and prevents common pitfalls
Use when creating GitHub Copilot instructions - provides repository-wide and path-specific formats, applyTo patterns, excludeAgent options, and natural language markdown style
Use when creating OpenCode agents - provides markdown format with YAML frontmatter, mode/tools/permission configuration, and best practices for specialized AI assistants
Comprehensive guide to troubleshooting Pulumi TypeScript errors, infrastructure issues, and best practices - covers common errors, Outputs handling, AWS Beanstalk deployment, and cost optimization
This is the meta skill for writing skills themselves.
This is a Claude skill for testing multi-package publishing.
Use when creating, improving, or troubleshooting Claude Code slash commands. Expert guidance on command structure, arguments, frontmatter, tool permissions, and best practices for building effective custom commands.
Use when creating Windsurf rules - provides plain markdown format with NO frontmatter, 12,000 character limit, and single-file structure requirements
Use when setting up npm publishing with GitHub Actions - provides trusted publishing with OIDC, provenance attestations, and monorepo configuration
Use when creating or improving Claude Code agents. Expert guidance on agent file structure, frontmatter, persona definition, tool access, model selection, and validation against schema.
Write content that sounds natural, conversational, and authentically human - avoiding AI-generated patterns, corporate speak, and generic phrasing
Use when deploying Node.js applications to AWS Elastic Beanstalk or troubleshooting deployment issues - provides dependency installation strategies, monorepo handling, and deployment best practices
Generate, edit, and beat-sync AI video with leading models in one workspace.
The world's fastest calendar for remote work
Transform Your Design with AI Designer by ImgCreator.ai
Revolutionizing Video Production with AI-Powered Creativity
Extend an image past the frame and let AI fill the new aspect ratio.
Discover your celebrity doppelgänger with StarByFace!
ChainClarity explains 700+ crypto whitepapers in plain English, with layered summaries, comparisons, research tools, alerts, and a $4.99 Pro plan.
Opus.ai: Revolutionize Your Web Experience