clawdskillz

clawdskillz

Engineering mindset skills for Claude Code. These skills activate different professional personas - skeptical engineers, pragmatic PMs, rigorous thinkers - to improve the quality of AI-assisted development.

Skills Overview

Skill	Phase	Purpose
self-doubt	Meta	Question your reasoning step-by-step
venture-capitalist	Pre-product	Validate business ideas, research competitors
product-manager	Planning	Scope features, clarify requirements
skeptic-engineer	Planning	Poke holes in technical approaches
test-engineer	Planning	Surface edge cases and scenarios
parallel-orchestrator	Plan→Impl	Break work into parallel subagent streams
framework-engineer	Implementation	Research libraries, find deprecations
brevity-engineer	Implementation	Ensure minimal, spec-aligned code
linting-engineer	Implementation	Run linters, enforce type consistency
test-writer	Implementation	Write fast, meaningful tests
documentation-engineer	Implementation	Maintain CLAUDE.md and docs
scaffold-engineer	Setup	Initialize projects with proper tooling

Installation

Option 1: Install via Marketplace (Recommended)

Add the marketplace to Claude Code:

/plugin marketplace add oxhagolli/clawdskillz

Then install the plugin set you want:

# Install all skills
/plugin install all-skills@clawdskillz

# Or install by category
/plugin install planning-skills@clawdskillz
/plugin install implementation-skills@clawdskillz
/plugin install setup-skills@clawdskillz
/plugin install thinking-skills@clawdskillz

Option 2: Install Directly

Install a specific plugin directly:

/plugin install all-skills@oxhagolli/clawdskillz

Option 3: Copy to Your Project

Copy the skills you want into your project's .claude/skills/ directory:

# Clone the repo
git clone https://github.com/oxhagolli/clawdskillz.git

# Copy specific skills
cp -r clawdskillz/skills/skeptic-engineer .claude/skills/
cp -r clawdskillz/skills/product-manager .claude/skills/

Option 4: Add to CLAUDE.md

Copy the content of any SKILL.md directly into your project's CLAUDE.md:

cat clawdskillz/skills/skeptic-engineer/SKILL.md >> CLAUDE.md

Plugin Sets

thinking-skills

• self-doubt - Rigorous step-by-step reasoning with self-questioning

planning-skills

• venture-capitalist - Business validation, competitor research, market analysis
• product-manager - Requirements, scope control, clarity
• skeptic-engineer - Technical approach critique, architecture risks
• test-engineer - Edge cases, scenarios to handle during planning
• parallel-orchestrator - Break approved plans into parallel subagent work streams

implementation-skills

• framework-engineer - Library selection, deprecation checks, best practices
• brevity-engineer - Code minimalism, repo standards, scope in code
• linting-engineer - Lint execution, type checking, type consistency
• test-writer - Integration tests, fast tests, pragmatic coverage
• documentation-engineer - CLAUDE.md per directory, ARCHITECTURE.md

setup-skills

• scaffold-engineer - Project initialization with pyscaffold, pre-commit

all-skills

All of the above.

Usage

Once installed, skills activate automatically based on context. You can also invoke them explicitly:

Apply the skeptic-engineer skill: I'm planning to rewrite our auth system to use JWTs...

Use self-doubt: Walk me through why this algorithm is O(n log n)...

Apply venture-capitalist: I want to build a new SaaS for...

Skill Separation

Each skill has a clear domain and explicitly defers to other skills:

Idea & Design:

venture-capitalist → "Should this exist?" (market, competition)
        ↓
product-manager → "What exactly should it do?" (scope, requirements)
        ↓
skeptic-engineer → "Will this approach work?" (architecture, risks)
        ↓
test-engineer → "What scenarios must we handle?" (edge cases)

Plan to Implementation:

parallel-orchestrator → "How do we parallelize this?" (subagent coordination)
        ↓
   ┌────┴────┐
   ▼         ▼
[Subagent] [Subagent]  → Each runs implementation skills
   └────┬────┘
        ▼
   Integration

Implementation (per subagent):

framework-engineer → "Which libraries should we use?" (dependencies)
        ↓
brevity-engineer → "Is this code minimal?" (scope in code)
        ↓
linting-engineer → "Does this pass checks?" (lint, types)
        ↓
test-writer → "Are tests meaningful and fast?" (test implementation)
        ↓
documentation-engineer → "Is this documented?" (CLAUDE.md, docs)

License

Apache License 2.0