{
  "slug": "karpathy-skills",
  "name": "karpathy-skills",
  "url": "https://github.com/multica-ai/andrej-karpathy-skills",
  "cat": "plan",
  "group": "claude",
  "tagline": "CLAUDE.md behavioral ruleset",
  "price": "Free",
  "tier": "free",
  "pick": true,
  "what": "A ~65-line CLAUDE.md built around four LLM coding failure modes Karpathy identified. Adds no tools — it changes how the agent decides: surface assumptions before coding, refuse speculative abstraction, keep diffs scoped, define a verifiable stopping condition. ~203k stars.",
  "why": "<b>Cheapest quality win on the entire list.</b> One file, no install, works alongside everything else. 'Surgical Changes' alone stops the drive-by refactors that break a working layout while fixing something unrelated.",
  "install": "curl -o CLAUDE.md .../main/CLAUDE.md",
  "category": {
    "key": "plan",
    "name": "Planning & reasoning",
    "desc": "Rulesets that make Claude think before it types. Cheapest quality win available.",
    "group": "claude",
    "groupName": "Claude Code toolkit"
  },
  "freshness": null,
  "health": {
    "ok": true,
    "status_code": 200,
    "final_url": null,
    "checked_at": "2026-08-18T20:13:57.562+00:00"
  },
  "corrected_at": null,
  "upstream_changed_at": null
}