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.
Install with Tessl CLI
npx tessl i github:forrestchang/andrej-karpathy-skills --skill karpathy-guidelines86
Does it follow best practices?
If you maintain this skill, you can automatically optimize it using the tessl CLI to improve its score:
npx tessl skill review --optimize ./path/to/skillAgent success when using this skill
Validation for skill structure
Discovery
67%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This description has a solid structure with explicit 'Use when...' guidance and addresses both what and when. However, the capabilities listed are somewhat abstract behavioral concepts rather than concrete actions, and the trigger terms could be expanded to capture more natural user language. The scope is broad enough that it may conflict with other coding-related skills.
Suggestions
Add more specific trigger terms users might naturally say, such as 'programming', 'debugging', 'best practices', 'clean code', or 'code quality'.
Make the capabilities more concrete by specifying what the guidelines actually do, e.g., 'Provides checklists for incremental changes, assumption documentation, and testable acceptance criteria'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (coding/LLM mistakes) and lists some actions ('avoid overcomplication, make surgical changes, surface assumptions, define verifiable success criteria'), but these are somewhat abstract behavioral concepts rather than concrete technical actions like 'extract text' or 'fill forms'. | 2 / 3 |
Completeness | Clearly answers both what ('Behavioral guidelines to reduce common LLM coding mistakes') and when ('Use when writing, reviewing, or refactoring code') with explicit trigger guidance using the 'Use when...' clause. | 3 / 3 |
Trigger Term Quality | Includes relevant keywords like 'writing', 'reviewing', 'refactoring', and 'code', but misses common variations users might say such as 'programming', 'debugging', 'fixing bugs', 'coding best practices', or 'clean code'. | 2 / 3 |
Distinctiveness Conflict Risk | The focus on 'LLM coding mistakes' provides some distinction, but 'writing, reviewing, or refactoring code' is broad enough to potentially conflict with other coding-related skills like code review tools, refactoring utilities, or general coding assistants. | 2 / 3 |
Total | 9 / 12 Passed |
Implementation
100%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is an exemplary skill file that demonstrates excellent conciseness while remaining highly actionable. The four sections are well-structured with clear headings, concrete examples, and explicit anti-patterns. The verification loop pattern in section 4 shows proper workflow thinking for a behavioral guideline skill.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every section is lean and purposeful. No explanation of concepts Claude already knows, no padding. The guidelines are stated directly with minimal but sufficient context. | 3 / 3 |
Actionability | Provides concrete, specific guidance with clear examples of what to do and not do. The transformation examples in section 4 ('Add validation' → 'Write tests...') are particularly actionable. The checklist format makes guidance immediately applicable. | 3 / 3 |
Workflow Clarity | Section 4 explicitly defines a verification loop pattern with clear checkpoints. The multi-step task template with verify steps demonstrates proper feedback loops. For a behavioral guideline skill, the workflow is appropriately clear. | 3 / 3 |
Progressive Disclosure | For a skill under 50 lines with no need for external references, the content is well-organized into four distinct, clearly-labeled sections. Each section is self-contained and easy to navigate. | 3 / 3 |
Total | 12 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Reviewed
Table of Contents
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.