MCP App StoreMCP App Store
from flock1,072
A set of coding principles prioritising simplicity, explicit assumptions, surgical edits, and verifiable success criteria to keep agent-produced code minimal an
Karpathy Coding Principles provides a compact playbook that guides an agent to write minimal, focused, and verifiable code. It instructs the agent to state assumptions, prefer the simplest solution that meets requirements, make only surgical changes to existing codebases, and define clear success criteria (tests or checks) before implementation. The guidance helps avoid speculative features, unnecessary abstractions, and unverified changes, improving code reviewability and reducing regressions.
Use this skill when the task involves writing or editing code where clarity, minimal surface area, and measurable outcomes matter — e.g., bug fixes, small feature additions, or test-driven fixes. Trigger it when the user requests concise solutions, explicit assumptions, or when touching existing code that should not be over-refactored. Ideal for senior-review-style edits where surgical changes are required.
Best applied to coding-capable agents (Copilot/Codex-style or Claude/Code-focused assistants) that can generate and reason about code changes and tests.
This skill has not been reviewed by our automated audit pipeline yet.
Run Execute
Orchestrates execution of work items across modes (autopilot, confirm, validate) with scripted init/complete tooling, plan/test/report artifacts and strict gati
Codex Collab
Bridge Claude and Codex to run tasks, perform code reviews, and run background jobs for research or implementation workflows.
AI Pair Collaboration
Orchestrates a small heterogeneous team of AI agents (author/developer + Codex reviewer + Gemini reviewer) to produce and review content or code in a coordinate
Crucible: Build
Orchestrates an end-to-end development pipeline: interactive design, planning, autonomous execution, and quality gates with TDD and iterative reviews.
Code Reviewer
Perform structured, prioritized code reviews that find correctness, security, performance, reliability, and testing issues and provide concrete fix suggestions.
Explorar base de código
Read and map a repository before making changes: discover structure, config, tests, patterns, and dependencies to avoid breaking conventions when modifying code
AIOS Test Architect & Quality Advisor
Advisory agent that reviews test architecture, creates quality gates, assesses risk, and suggests test strategies and improvements.
Collaborating with Claude Code
Bridge to delegate code implementation, review, and tests to Claude Code via a JSON bridge script that returns structured output for multi-turn sessions.