karpathy-guidelines

Behavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.

1.14x

Quality

88%

Does it follow best practices?

Impact

92%

1.14x

Average score across 3 eval scenarios

Securityby

Passed

No findings from the security scan

Quality

Content

100%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A tight, well-structured behavioral-guidelines skill that assumes Claude's competence, gives concrete directive guidance with worked examples, and sequences its guidance with explicit verification checkpoints. No bundle files are present or needed, and the single-file organization is appropriate.

Dimension	Reasoning	Score
Conciseness	Lean directive bullets with no padding or explanation of concepts Claude already knows; every line (including the 'Tradeoff' note) earns its place.	3 / 3
Actionability	Concrete directives ('No abstractions for single-use code', 'Match existing style') plus worked transformation examples ('Add validation' → 'Write tests for invalid inputs, then make them pass') and a copy-ready plan template; actionable despite being instruction-only.	3 / 3
Workflow Clarity	Four sequenced sections form a clear flow, and the Goal-Driven section provides explicit 'verify: [check]' checkpoints plus a 'Loop until verified' feedback loop.	3 / 3
Progressive Disclosure	A single, well-organized SKILL.md with clearly headed sections and no nested external references; appropriate for a self-contained behavioral skill.	3 / 3
	Total	12 / 12 Passed

Description

77%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A well-structured description that clearly states purpose and an explicit 'Use when' trigger with several concrete capabilities. It loses points on trigger-term naturalness and distinctiveness because the trigger phrasing is broad and leans on internal jargon rather than varied user utterances.

Suggestions

Add more natural user phrasings to the trigger (e.g., 'when the user asks to write, review, refactor, or simplify code') to broaden trigger-term coverage beyond the current jargon.

Sharpen the trigger to distinguish this skill from generic coding assistance, e.g., 'Use when writing or refactoring code where over-engineering, scope creep, or unclear assumptions are a risk.'

Consider trimming internal vocabulary like 'surgical changes' and 'verifiable success criteria' from the trigger clause in favor of terms a user would actually say.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria' — going beyond a domain label to specific capabilities.	3 / 3
Completeness	Explicitly answers both what ('Behavioral guidelines to reduce common LLM coding mistakes') and when ('Use when writing, reviewing, or refactoring code').	3 / 3
Trigger Term Quality	'writing, reviewing, or refactoring code' are natural user terms, but the bulk of the phrasing ('surgical changes', 'surface assumptions', 'verifiable success criteria') is capability-jargon users would not naturally say, leaving common variations under-covered.	2 / 3
Distinctiveness Conflict Risk	The anti-overcomplication niche is distinct, but the trigger 'writing, reviewing, or refactoring code' is broad and would overlap with many general coding skills.	2 / 3
	Total	10 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: OpenBMB/PilotDeck
Path: skills/karpathy-guidelines/SKILL.md
Commit: c88d7d1

Reviewed: about 14 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.