[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"extension-marketplace-forrestchang-andrej-karpathy-skills-en":3,"guides-for-forrestchang-andrej-karpathy-skills":312,"similar-k174fg5qph3f55akvm05r7ys2d86mxkh-en":313},{"_creationTime":4,"_id":5,"children":6,"community":49,"display":50,"evaluation":55,"identity":252,"isFallback":238,"parentExtension":254,"providers":255,"relations":262,"repo":263,"tags":309,"workflow":310},1778687058458.6985,"k174fg5qph3f55akvm05r7ys2d86mxkh",[7],{"_creationTime":8,"_id":9,"community":10,"display":12,"identity":17,"providers":22,"relations":42,"tags":44,"workflow":45},1778687058458.6987,"k1739sx217s32es9e9p803xn1h86m6zv",{"reviewCount":11},0,{"description":13,"installMethods":14,"name":15,"sourceUrl":16},"Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations on LLM coding pitfalls",{"claudeCode":15},"andrej-karpathy-skills","https://github.com/forrestchang/andrej-karpathy-skills",{"basePath":18,"githubOwner":19,"githubRepo":15,"locale":20,"slug":15,"type":21},"","forrestchang","en","plugin",{"evaluate":23,"extract":36},{"promptVersionExtension":24,"promptVersionScoring":25,"score":26,"tags":27,"targetMarket":34,"tier":35},"3.0.0","4.4.0",99,[28,29,30,31,32,33],"guidelines","llm","coding-practices","refactoring","documentation","karpathy","global","verified",{"commitSha":37,"license":38,"plugin":39},"HEAD","MIT",{"mcpCount":11,"provider":40,"skillCount":41},"classify",1,{"parentExtensionId":5,"repoId":43},"kd7awe12sdp9xgt83qhzh8se5186mz5x",[30,32,28,33,29,31],{"evaluatedAt":46,"extractAt":47,"updatedAt":48},1778687139483,1778687058458,1778687230225,{"reviewCount":11},{"description":51,"installMethods":52,"name":54,"sourceUrl":16},"Behavioral guidelines to reduce common LLM coding mistakes, derived from Andrej Karpathy's observations",{"claudeCode":53},"forrestchang/andrej-karpathy-skills","Karpathy-Inspired Claude Code Guidelines",{"_creationTime":56,"_id":57,"extensionId":5,"locale":20,"result":58,"trustSignals":236,"workflow":250},1778687113381.9272,"kn7985bf2zbzy3yz8krx0cw0k586nteh",{"checks":59,"evaluatedAt":205,"extensionSummary":206,"features":207,"nonGoals":212,"practices":216,"prerequisites":220,"promptVersionExtension":221,"promptVersionScoring":25,"purpose":222,"rationale":223,"score":224,"summary":225,"tags":226,"targetMarket":34,"tier":35,"useCases":231,"workflow":235},[60,65,68,71,75,78,83,87,90,93,97,101,104,108,111,114,117,120,123,126,130,134,138,142,146,149,152,155,159,162,165,168,171,174,177,181,185,189,192,196,199,202],{"category":61,"check":62,"severity":63,"summary":64},"Practical Utility","Problem relevance","pass","The description clearly names the problem of \"common LLM coding mistakes\" and the pain point of \"wrong assumptions, hidden confusion, missing tradeoffs, overcomplication, and orthogonal edits.\"",{"category":61,"check":66,"severity":63,"summary":67},"Unique selling proposition","The extension provides four specific, actionable principles that steer LLM behavior beyond default prompts, directly addressing named pitfalls with a structured approach.",{"category":61,"check":69,"severity":63,"summary":70},"Production readiness","The extension is ready for production use, offering a CLAUDE.md file for per-project integration or a Claude Code plugin for broader availability, covering the entire lifecycle of guideline application.",{"category":72,"check":73,"severity":63,"summary":74},"Scope","Single responsibility principle","The extension focuses solely on providing behavioral guidelines for LLM coding practices, aligning with its stated purpose and not extending into unrelated domains.",{"category":72,"check":76,"severity":63,"summary":77},"Description quality","The displayed description accurately and concisely reflects the extension's purpose and derivation from Andrej Karpathy's observations on LLM coding pitfalls.",{"category":79,"check":80,"severity":81,"summary":82},"Invocation","Scoped tools","not_applicable","This extension is a marketplace catalog and does not expose specific tools or commands.",{"category":84,"check":85,"severity":81,"summary":86},"Documentation","Configuration & parameter reference","As a marketplace, this extension does not have configurable parameters or options requiring documentation.",{"category":72,"check":88,"severity":81,"summary":89},"Tool naming","This extension does not expose user-facing tools or commands.",{"category":72,"check":91,"severity":81,"summary":92},"Minimal I/O surface","This extension does not expose tools or commands with input/output schemas.",{"category":94,"check":95,"severity":63,"summary":96},"License","License usability","The extension is licensed under MIT, clearly stated in the README file.",{"category":98,"check":99,"severity":63,"summary":100},"Maintenance","Commit recency","The last commit was on 2026-04-20, which is within the last 3 months.",{"category":98,"check":102,"severity":81,"summary":103},"Dependency Management","The extension does not appear to use any third-party dependencies.",{"category":105,"check":106,"severity":81,"summary":107},"Security","Secret Management","The extension does not handle or expose any secrets.",{"category":105,"check":109,"severity":81,"summary":110},"Injection","The extension does not load or process third-party data that could be subject to injection attacks.",{"category":105,"check":112,"severity":81,"summary":113},"Transitive Supply-Chain Grenades","The extension does not load or include external files at runtime.",{"category":105,"check":115,"severity":81,"summary":116},"Sandbox Isolation","The extension is a marketplace and does not execute code or modify files.",{"category":105,"check":118,"severity":81,"summary":119},"Sandbox escape primitives","The extension does not execute code and thus cannot have sandbox escape primitives.",{"category":105,"check":121,"severity":81,"summary":122},"Data Exfiltration","The extension does not perform any outbound calls or handle confidential data.",{"category":105,"check":124,"severity":63,"summary":125},"Hidden Text Tricks","The bundled content (README.md, CLAUDE.md) appears free of hidden-steering tricks, control characters, or invisible Unicode tags.",{"category":127,"check":128,"severity":81,"summary":129},"Hooks","Opaque code execution","The extension does not contain any scripts or hooks that involve opaque code execution.",{"category":131,"check":132,"severity":81,"summary":133},"Portability","Structural Assumption","This extension does not make structural assumptions about user project organization as it provides guidelines, not code.",{"category":135,"check":136,"severity":63,"summary":137},"Trust","Issues Attention","There are 0 open and 0 closed issues in the last 90 days, indicating no current attention needed or a new/inactive project state, which is acceptable.",{"category":139,"check":140,"severity":63,"summary":141},"Versioning","Release Management","The extension declares version 1.0.0 in its marketplace metadata and README, indicating a meaningful version is present.",{"category":143,"check":144,"severity":81,"summary":145},"Code Execution","Validation","The extension does not involve executable code or structured output that requires validation.",{"category":105,"check":147,"severity":81,"summary":148},"Unguarded Destructive Operations","The extension is a marketplace and does not perform any destructive operations.",{"category":143,"check":150,"severity":81,"summary":151},"Error Handling","This extension does not execute code and therefore has no error paths visible to the user.",{"category":143,"check":153,"severity":81,"summary":154},"Logging","The extension is a marketplace and does not perform actions that require local audit logging.",{"category":156,"check":157,"severity":81,"summary":158},"Compliance","GDPR","The extension does not operate on personal data.",{"category":156,"check":160,"severity":63,"summary":161},"Target market","The extension's guidelines are language-agnostic and do not contain regional signals, making it globally applicable.",{"category":131,"check":163,"severity":81,"summary":164},"Runtime stability","The extension does not have runtime dependencies on specific editors, shells, or operating systems.",{"category":84,"check":166,"severity":63,"summary":167},"README","The README file clearly states the extension's purpose and provides installation instructions.",{"category":72,"check":169,"severity":81,"summary":170},"Tool surface size","This is a marketplace extension and does not expose tools or commands.",{"category":79,"check":172,"severity":81,"summary":173},"Overlapping near-synonym tools","This extension does not expose tools or commands.",{"category":84,"check":175,"severity":63,"summary":176},"Phantom features","All features mentioned in the README and marketplace metadata, such as the four principles and installation methods, are implemented or present.",{"category":178,"check":179,"severity":63,"summary":180},"Install","Installation instruction","The README provides clear installation instructions for both the Claude Code plugin and the CLAUDE.md file, including copy-pasteable examples.",{"category":182,"check":183,"severity":81,"summary":184},"Errors","Actionable error messages","As this extension does not have executable code or user-facing error paths, this check is not applicable.",{"category":186,"check":187,"severity":81,"summary":188},"Execution","Pinned dependencies","The extension does not use third-party dependencies that require pinning.",{"category":72,"check":190,"severity":81,"summary":191},"Dry-run preview","The extension does not involve state-changing operations or outbound data transmission that would require a dry-run preview.",{"category":193,"check":194,"severity":81,"summary":195},"Protocol","Idempotent retry & timeouts","This extension does not involve remote calls or state-changing operations.",{"category":156,"check":197,"severity":81,"summary":198},"Telemetry opt-in","The extension does not emit any telemetry.",{"category":72,"check":200,"severity":63,"summary":201},"Theme declaration","The README clearly declares a curation theme focused on improving LLM coding behavior by addressing common mistakes, derived from Andrej Karpathy's observations.",{"category":139,"check":203,"severity":63,"summary":204},"Per-entry version metadata","The marketplace metadata correctly pins the plugin to a specific version (1.0.0) rather than a mutable Git ref like 'main'.",1778687111932,"This marketplace extension curates behavioral guidelines for AI coding assistants, aimed at mitigating common mistakes like incorrect assumptions, overcomplication, and unnecessary code modifications. It provides these guidelines in a single CLAUDE.md file or as a Claude Code plugin.",[208,209,210,211],"Actionable coding guidelines derived from expert observations","Addresses common LLM pitfalls: incorrect assumptions, overcomplication, orthogonal edits","Provides four core principles: Think Before Coding, Simplicity First, Surgical Changes, Goal-Driven Execution","Offers integration via Claude Code plugin or CLAUDE.md file",[213,214,215],"Providing specific code implementations for tasks.","Acting as a general-purpose code formatter or linter.","Replacing the need for human code review entirely.",[217,218,219],"Coding quality","LLM behavior","Prompt engineering",[],"3.1.0","To help users improve the quality and focus of AI-generated code by applying established principles that prevent common LLM coding errors.","Excellent documentation and clear problem definition; all checks passed or were not applicable.",98,"A high-quality marketplace extension providing clear, actionable LLM coding guidelines derived from expert observations.",[29,227,228,229,230],"coding-guidelines","best-practices","claude-code","prompt-engineering",[232,233,234],"Use when AI-generated code frequently contains errors or unexpected changes.","Incorporate into projects to enforce consistent coding quality from AI assistants.","Leverage to guide AI assistants toward more focused and verifiable task execution.",[],{"codeQuality":237,"collectedAt":239,"documentation":240,"maintenance":243,"security":248,"testCoverage":249},{"hasLockfile":238},false,1778687060055,{"descriptionLength":241,"readmeSize":242},103,6198,{"closedIssues90d":11,"forks":244,"hasChangelog":238,"manifestVersion":245,"openIssues90d":11,"pushedAt":246,"stars":247},13000,"1.0.0",1776679504000,128093,{"hasNpmPackage":238,"smitheryVerified":238},{"hasCi":238,"hasTests":238},{"updatedAt":251},1778687113381,{"basePath":18,"githubOwner":19,"githubRepo":15,"locale":20,"slug":15,"type":253},"marketplace",null,{"evaluate":256,"extract":258},{"promptVersionExtension":221,"promptVersionScoring":25,"score":224,"tags":257,"targetMarket":34,"tier":35},[29,227,228,229,230],{"commitSha":37,"license":38,"marketplace":259,"plugin":261},{"name":260,"pluginCount":41},"karpathy-skills",{"mcpCount":11,"provider":40,"skillCount":11},{"repoId":43},{"_creationTime":264,"_id":43,"identity":265,"providers":266,"workflow":305},1778687054482.7363,{"githubOwner":19,"githubRepo":15,"sourceUrl":16},{"classify":267,"discover":299,"github":302},{"commitSha":37,"extensions":268},[269,279,289],{"basePath":18,"description":51,"displayName":260,"installMethods":270,"rationale":271,"selectedPaths":272,"source":278,"sourceLanguage":20,"type":253},{"claudeCode":53},"marketplace.json at .claude-plugin/marketplace.json",[273,276],{"path":274,"priority":275},".claude-plugin/marketplace.json","mandatory",{"path":277,"priority":275},"README.md","rule",{"basePath":18,"description":13,"displayName":15,"installMethods":280,"license":38,"rationale":281,"selectedPaths":282,"source":278,"sourceLanguage":20,"type":21},{"claudeCode":15},"plugin manifest at .claude-plugin/plugin.json",[283,285,286],{"path":284,"priority":275},".claude-plugin/plugin.json",{"path":277,"priority":275},{"path":287,"priority":288},"skills/karpathy-guidelines/SKILL.md","medium",{"basePath":290,"description":291,"displayName":292,"installMethods":293,"rationale":294,"selectedPaths":295,"source":278,"sourceLanguage":20,"type":298},"skills/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.","karpathy-guidelines",{"claudeCode":53},"SKILL.md frontmatter at skills/karpathy-guidelines/SKILL.md",[296],{"path":297,"priority":275},"SKILL.md","skill",{"sources":300},[301],"manual",{"closedIssues90d":11,"description":303,"forks":244,"openIssues90d":11,"pushedAt":246,"readmeSize":242,"stars":247,"topics":304},"A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.",[],{"classifiedAt":306,"discoverAt":307,"extractAt":308,"githubAt":308,"updatedAt":306},1778687058299,1778687054482,1778687056438,[228,229,227,29,230],{"evaluatedAt":251,"extractAt":47,"updatedAt":311},1778687230040,[],[314,344,373,405,435],{"_creationTime":315,"_id":316,"community":317,"display":318,"identity":324,"providers":329,"relations":338,"tags":340,"workflow":341},1778691765524.3726,"k17c1ndsb7hncr0k5vadc69rnx86njvh",{"reviewCount":11},{"description":319,"installMethods":320,"name":322,"sourceUrl":323},"🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman",{"claudeCode":321},"juliusbrussee/caveman","caveman-repo","https://github.com/juliusbrussee/caveman",{"basePath":325,"githubOwner":326,"githubRepo":327,"locale":20,"slug":328,"type":253},".agents/plugins","juliusbrussee","caveman","plugins",{"evaluate":330,"extract":336},{"promptVersionExtension":221,"promptVersionScoring":25,"score":331,"tags":332,"targetMarket":34,"tier":35},100,[333,334,335,229,230],"token-compression","llm-efficiency","cli",{"commitSha":37,"marketplace":337},{"name":322,"pluginCount":11},{"repoId":339},"kd7890f0sfy0ymf1pakpx7vqch86nmew",[229,335,334,230,333],{"evaluatedAt":342,"extractAt":343,"updatedAt":342},1778691782274,1778691765524,{"_creationTime":345,"_id":346,"community":347,"display":348,"identity":354,"providers":356,"relations":366,"tags":368,"workflow":369},1778696712851.9736,"k1730g2nnwga3htdfycmc8h1gd86m3cs",{"reviewCount":11},{"description":349,"installMethods":350,"name":352,"sourceUrl":353},"2x the effective context with smart weight-loss for Claude Code — prune bloated sessions, protect agent teams from compaction, monitor token usage with MCP tools",{"claudeCode":351},"Ruya-AI/cozempic","cozempic","https://github.com/Ruya-AI/cozempic",{"basePath":18,"githubOwner":355,"githubRepo":352,"locale":20,"slug":352,"type":253},"Ruya-AI",{"evaluate":357,"extract":363},{"promptVersionExtension":221,"promptVersionScoring":25,"score":331,"tags":358,"targetMarket":34,"tier":35},[359,360,229,361,362],"context","optimization","session-management","ai-productivity",{"commitSha":37,"marketplace":364,"plugin":365},{"name":352,"pluginCount":41},{"mcpCount":11,"provider":40,"skillCount":11},{"repoId":367},"kd79d77qmyh3826dwhk7ynx9xd86nmqm",[362,229,359,360,361],{"evaluatedAt":370,"extractAt":371,"updatedAt":372},1778696729067,1778696712852,1778696924064,{"_creationTime":374,"_id":375,"community":376,"display":377,"identity":383,"providers":386,"relations":399,"tags":401,"workflow":402},1778683562157.875,"k174pnm5ch9ab6fr1etef2f2b586m74b",{"reviewCount":11},{"description":378,"installMethods":379,"name":381,"sourceUrl":382},"Persistent memory and cognitive profiling plugins for Claude Code",{"claudeCode":380},"cdeust/Cortex","cortex-plugins","https://github.com/cdeust/Cortex",{"basePath":18,"githubOwner":384,"githubRepo":385,"locale":20,"slug":385,"type":253},"cdeust","Cortex",{"evaluate":387,"extract":396},{"promptVersionExtension":221,"promptVersionScoring":25,"score":331,"tags":388,"targetMarket":34,"tier":35},[389,390,391,229,392,393,394,395],"memory","cognitive-profiling","mcp","knowledge-graph","codebase-analysis","postgresql","pgvector",{"commitSha":37,"marketplace":397,"plugin":398},{"name":381,"pluginCount":41},{"mcpCount":11,"provider":40,"skillCount":11},{"repoId":400},"kd79gxpemvkr09a7zsb3h8kmah86nvgf",[229,393,390,392,391,389,395,394],{"evaluatedAt":403,"extractAt":404,"updatedAt":403},1778683583007,1778683562157,{"_creationTime":406,"_id":407,"community":408,"display":409,"identity":415,"providers":417,"relations":428,"tags":430,"workflow":431},1778697668802.9524,"k1746m1dnjtegrp9ftydmb5wn186mxsv",{"reviewCount":11},{"description":410,"installMethods":411,"name":413,"sourceUrl":414},"One-stop shop for building AI-powered products and businesses with Stripe.",{"claudeCode":412},"stripe/agent-toolkit","stripe","https://github.com/stripe/agent-toolkit",{"basePath":18,"githubOwner":413,"githubRepo":416,"locale":20,"slug":416,"type":253},"agent-toolkit",{"evaluate":418,"extract":425},{"promptVersionExtension":221,"promptVersionScoring":25,"score":331,"tags":419,"targetMarket":34,"tier":35},[413,420,421,422,423,29,424],"ai","sdk","typescript","python","agent",{"commitSha":37,"marketplace":426,"plugin":427},{"name":413,"pluginCount":41},{"mcpCount":11,"provider":40,"skillCount":11},{"repoId":429},"kd74b9k48z93rf42bgdvfdcahh86nzpd",[424,420,29,423,421,413,422],{"evaluatedAt":432,"extractAt":433,"updatedAt":434},1778697682749,1778697668803,1778697814342,{"_creationTime":436,"_id":437,"community":438,"display":439,"identity":445,"providers":447,"relations":455,"tags":457,"workflow":458},1778693511416.3645,"k17de181ddh1gjh0zdntr3qxyn86nqna",{"reviewCount":11},{"description":440,"installMethods":441,"name":443,"sourceUrl":444},"Context window optimization for AI coding agents. Sandboxes tool output, 98% reduction. 15 platforms",{"claudeCode":442},"mksglu/context-mode","context-mode","https://github.com/mksglu/context-mode",{"basePath":325,"githubOwner":446,"githubRepo":443,"locale":20,"slug":328,"type":253},"mksglu",{"evaluate":448,"extract":453},{"promptVersionExtension":221,"promptVersionScoring":25,"score":331,"tags":449,"targetMarket":34,"tier":35},[359,360,29,450,451,452],"productivity","code-generation","sandbox",{"commitSha":37,"marketplace":454},{"name":443,"pluginCount":11},{"repoId":456},"kd764b2fctbqg4b8j8y6xvmkvs86m29m",[451,359,29,360,450,452],{"evaluatedAt":459,"extractAt":460,"updatedAt":461},1778693535406,1778693511416,1778693816674]