Karpathy Coder
Plugin Verifiziert AktivActive coding discipline enforcer based on Karpathy's 4 principles: surface assumptions, keep it simple, make surgical changes, define verifiable goals. Ships 4 Python tools (complexity_checker, diff_surgeon, assumption_linter, goal_verifier), a review agent, /karpathy-check slash command, and a pre-commit hook. All tools stdlib-only.
To actively enforce rigorous coding discipline and prevent common LLM-generated code pitfalls by providing concrete tools and hooks for analysis and review.
Funktionen
- Enforces Karpathy's 4 coding principles
- Includes 4 stdlib-only Python analysis tools
- Provides a review agent and slash command
- Integrates a non-blocking pre-commit hook
- Offers detailed reports with actionable recommendations
Anwendungsfälle
- Before committing code to catch noise and overcomplication early
- After completing a feature as a sanity check before a PR
- When suspecting LLM-generated code is overcomplicated or lacks verification
- To enforce coding standards within a development team
Nicht-Ziele
- Performing automated code refactoring or fixes
- Replacing formal code review processes entirely
- Enforcing stylistic conventions not related to Karpathy's principles
Installation
Zuerst Marketplace hinzufügen
/plugin marketplace add alirezarezvani/claude-skills/plugin install karpathy-coder@claude-code-skillsQualitätspunktzahl
VerifiziertVertrauenssignale
Ähnliche Erweiterungen
Autoresearch Agent
100Autonomous experiment loop that optimizes any file by a measurable metric. 5 slash commands, 8 evaluators, configurable loop intervals (10min to monthly).
Context7 Plugin
100Upstash Context7 MCP-Server für die Abfrage aktueller Dokumentationen. Ruft versionsspezifische Dokumentationen und Codebeispiele direkt aus Quell-Repositories in den LLM-Kontext.
Dotforge Stack Python Fastapi
100Python 3.12+ with FastAPI, async/await, type hints, and Ruff linting rules for Claude Code.
Skill Optimizer
99Analysiert und optimiert Ihre Agent Skills (SKILL.md) mithilfe von Sitzungsdaten und forschungsgestützten statischen Prüfungen. Funktioniert mit Claude Code, Codex und jedem Agent Skills-kompatiblen Agenten.
Plannotator Copilot
99Interactive Plan & Code Review for GitHub Copilot CLI. Visual annotations, team sharing, structured feedback.
Python Development
98Modern Python development with Python 3.12+, Django, FastAPI, async patterns, and production best practices