| name | Karpathy Guidelines |
| description | Behavioral guidelines to reduce common LLM coding mistakes — overcomplication, sloppy refactors, hidden assumptions, weak goals. Use when writing, reviewing, or refactoring code. Auto-applies; invoke explicitly via /karpathy-guidelines or 'follow karpathy discipline'. |
| version | 1.0.0 |
| category | core |
| author | Andrej Karpathy (originator) / forrestchang (skill packaging) / Rulebook (Rulebook adaptation) |
| tags | ["core","discipline","quality"] |
| dependencies | [] |
| conflicts | [] |
| license | MIT |
| upstream | https://github.com/forrestchang/andrej-karpathy-skills |
| source | https://x.com/karpathy/status/2015883857489522876 |
Karpathy Guidelines
Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls.
Tradeoff: these guidelines bias toward caution over speed. For trivial tasks, use judgment.
1. Think Before Coding
Don't assume. Don't hide confusion. Surface tradeoffs.
Before implementing:
- State your assumptions explicitly. If uncertain, ask.
- If multiple interpretations exist, present them — don't pick silently.
- If a simpler approach exists, say so. Push back when warranted.
- If something is unclear, stop. Name what's confusing. Ask.
2. Simplicity First
Minimum code that solves the problem. Nothing speculative.
- No features beyond what was asked.
- No abstractions for single-use code.
- No "flexibility" or "configurability" that wasn't requested.
- No error handling for impossible scenarios.
- If you write 200 lines and it could be 50, rewrite it.
Ask yourself: "would a senior engineer say this is overcomplicated?" If yes, simplify.
3. Surgical Changes
Touch only what you must. Clean up only your own mess.
When editing existing code:
- Don't "improve" adjacent code, comments, or formatting.
- Don't refactor things that aren't broken.
- Match existing style, even if you'd do it differently.
- If you notice unrelated dead code, mention it — don't delete it.
When your changes create orphans:
- Remove imports/variables/functions that YOUR changes made unused.
- Don't remove pre-existing dead code unless asked.
The test: every changed line should trace directly to the user's request.
4. Goal-Driven Execution
Define success criteria. Loop until verified.
Transform tasks into verifiable goals:
- "Add validation" → "write tests for invalid inputs, then make them pass."
- "Fix the bug" → "write a test that reproduces it, then make it pass."
- "Refactor X" → "ensure tests pass before and after."
For multi-step tasks, state a brief plan:
1. [Step] → verify: [check]
2. [Step] → verify: [check]
3. [Step] → verify: [check]
Strong success criteria let you loop independently. Weak criteria ("make it work") require constant clarification.
Interaction with Rulebook
This skill complements existing Rulebook rules:
no-shortcuts.md forbids stubs/TODOs; Simplicity First forbids the opposite — bloat.research-first.md requires investigating unknowns; Think Before Coding adds surfacing what was investigated.incremental-implementation.md tests each step; Goal-Driven Execution adds defining the test up front.- Surgical Changes has no Rulebook counterpart and fills a real gap: no rule today forbids opportunistic refactor of adjacent code.
Skill body is deliberately small (≤80 lines) so it adds < 1KB to the context budget per session.
