LearnPrompt/andrej-karpathy-skills
Skill Claude CodeCodex
Apply the Karpathy understanding-first principle — verify what AI built, own your understanding, map the jagged capability landscape of LLMs. Use this skill when the user has used AI to build something and needs to verify it, wants to audit their own understanding, needs to know where LLMs are reliable vs unreliable…
LearnPrompt/andrej-karpathy-skills
Skill Claude CodeCodex
Transition from vibe coding (fast prototyping) to agentic engineering (rigorous iteration). Use this skill when the user has a rough prototype and wants to level it up, needs to go from "it kind of works" to production-quality, wants to understand the difference between vibe coding and agentic coding, or says "polish…
Skill Claude CodeCodex
Audit and govern Skills, plugins, and MCP servers across Codex, Claude Code, and compatible Agent hosts. Use when the user asks to inventory, deduplicate, classify, scope, archive, or reduce the context cost of Agent capabilities; asks why there are so many Skills; wants global/project/trigger decisions; or says…
Skill Claude CodeCodex
搭子.skill / Partner — two-direction cost-split workflow between Claude Code and Codex. Direction A (Codex-driven): Codex orchestrates and implements; Claude Code plans (/goal), polishes UI/interaction, and runs final /codex:review in one reused session. Direction B (Claude-driven): Claude plans and splits the work…
Skill Claude CodeCodex
点子王 (Idea King) — reason from first principles & run adversarial review. A thinking partner for the Partner (搭子) workflow and for standalone use. Use when the user says "点子王", "idea king", "第一性原理", "从第一性原理出发", "奥卡姆剃刀", "墨菲定律", "科斯定理", "对抗式审查", "挑战这个方案", "attack this plan", "盘问", "grill this plan", or when a plan /…
Plugin Claude Code
Plugin marketplace listing 1 plugin: loop-engineering.
Skill Claude CodeCodex
A planning framework for turning a vague request for repeated automation into a self-running workflow, called a loop. It defines the goal, triggers, task roles, connections, saved state, checks, and safety rules before producing commands for Claude Code or Codex.
Skill Claude CodeCodex
One source of truth for local AI agent skills: audit Codex, Claude, Hermes, OpenClaw, OpenCode, workspace skills, and shared roots; score hygiene, diff conflicting installs, classify drift as pristine/dirty/local-only against recorded baselines, deduplicate into one canonical source with symlinks, and migrate the same…