{"id":"d0c086a8-cf8e-46a2-8c27-6012c3b54e41","entityType":"agent","slug":"clawhub-zhaobod1-huo15-karpathy-guidelines","name":"Huo15 Karpathy Guidelines","canonicalUrl":"https://www.xpersona.co/agent/clawhub-zhaobod1-huo15-karpathy-guidelines","canonicalPath":"/agent/clawhub-zhaobod1-huo15-karpathy-guidelines","generatedAt":"2026-10-11T07:43:50.845Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-11T04:03:43.294Z","emptyReason":null},"description":"将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill Skill: Huo15 Karpathy Guidelines Owner: zhaobod1 Summary: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill Tags: latest:1.0.3 Version history: v1.0.3 | 2026-04-24T04:15:19.827Z | auto - No file changes detected in this version. - Functionality, features, and documentation remain unchanged from the previous release. v1.0.2 | 2026-04-23T03:21:44.545Z | user v1.0.2 把本地工作态同步到 clawhub（之前本地版本号落后于 clawhub） v1.0.1 | 2026-0","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1.2K downloads reported by the source. Last updated 10/11/2026.","installCommand":"clawhub skill install s17f6q72skfgyjycm2frgdc8dn83v5mk:huo15-karpathy-guidelines","sourceUrl":"https://clawhub.ai/zhaobod1/huo15-karpathy-guidelines","homepage":"https://clawhub.ai/zhaobod1/skills/huo15-karpathy-guidelines","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/zhaobod1/huo15-karpathy-guidelines","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/zhaobod1/skills/huo15-karpathy-guidelines","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":61,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill Skill: Huo15 Karpathy Guidelines Owner: zhaobod1 Summary: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill "},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-11T04:03:43.294Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T04:03:43.294Z","emptyReason":null},"stars":null,"forks":null,"downloads":1166,"packageName":null,"latestVersion":"1.0.3","tractionLabel":"1.2K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-11T04:03:43.287Z","emptyReason":null},"lastUpdatedAt":"2026-10-11T04:03:43.294Z","lastCrawledAt":"2026-10-11T04:03:43.287Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-12T04:03:43.287Z","lastVerifiedAt":null,"highlights":[{"version":"1.0.3","createdAt":"2026-04-24T04:15:19.827Z","changelog":"- No file changes detected in this version. - Functionality, features, and documentation remain unchanged from the previous release.","fileCount":4,"zipByteSize":4447},{"version":"1.0.2","createdAt":"2026-04-23T03:21:44.545Z","changelog":"v1.0.2 把本地工作态同步到 clawhub（之前本地版本号落后于 clawhub）","fileCount":3,"zipByteSize":3432},{"version":"1.0.1","createdAt":"2026-04-21T17:08:36.647Z","changelog":"- Added a frontmatter block with name, version, description, and aliases to SKILL.md for improved metadata. - No changes to functionality or guideline content; documentation only.","fileCount":3,"zipByteSize":3430},{"version":"1.0.0","createdAt":"2026-04-21T16:56:55.855Z","changelog":"- Initial release packaging Karpathy's four key LLM coding guidelines as an OpenClaw Skill. - Helps improve code quality and avoid common LLM coding pitfalls. - Supports multiple triggers including “karpathy”, “行为规范”, and more. - Summarizes best practices: ask before guessing, prioritize simplicity, make targeted changes, and execute with clear goals. - Includes ready-to-use usage instructions and a quick reference script.","fileCount":3,"zipByteSize":3368}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17f6q72skfgyjycm2frgdc8dn83v5mk:huo15-karpathy-guidelines","setupComplexity":"low","setupSteps":["Setup complexity is LOW. This package is likely designed for quick installation with minimal external side-effects.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-11T07:43:50.845Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-zhaobod1-huo15-karpathy-guidelines/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"high","updatedAt":"2026-10-11T04:03:43.294Z","emptyReason":null},"readme":"Skill: Huo15 Karpathy Guidelines\n\nOwner: zhaobod1\n\nSummary: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill\n\nTags: latest:1.0.3\n\nVersion history:\n\nv1.0.3 | 2026-04-24T04:15:19.827Z | auto\n\n- No file changes detected in this version.\n- Functionality, features, and documentation remain unchanged from the previous release.\n\nv1.0.2 | 2026-04-23T03:21:44.545Z | user\n\nv1.0.2 把本地工作态同步到 clawhub（之前本地版本号落后于 clawhub）\n\nv1.0.1 | 2026-04-21T17:08:36.647Z | auto\n\n- Added a frontmatter block with name, version, description, and aliases to SKILL.md for improved metadata.\n- No changes to functionality or guideline content; documentation only.\n\nv1.0.0 | 2026-04-21T16:56:55.855Z | auto\n\n- Initial release packaging Karpathy's four key LLM coding guidelines as an OpenClaw Skill.\n- Helps improve code quality and avoid common LLM coding pitfalls.\n- Supports multiple triggers including “karpathy”, “行为规范”, and more.\n- Summarizes best practices: ask before guessing, prioritize simplicity, make targeted changes, and execute with clear goals.\n- Includes ready-to-use usage instructions and a quick reference script.\n\nArchive index:\n\nArchive v1.0.3: 4 files, 4447 bytes\n\nFiles: _meta.json (144b), scripts/karpathy.sh (1903b), skill-card.md (1745b), SKILL.md (4007b)\n\nFile v1.0.3:SKILL.md\n\n---\nname: huo15-karpathy-guidelines\nversion: 1.0.2\ndescription: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill\naliases:\n  - 火一五卡帕西准则\n  - 火一五行为准则\n  - Karpathy准则\n---\n\n# SKILL.md — huo15-karpathy-guidelines\n\n## Name\n\nhuo15-karpathy-guidelines\n\n## Description\n\nKarpathy 行为准则技能 — 将 Andrej Karpathy 的 LLM 编程四大行为规范（71K⭐）打包为 OpenClaw Skill，帮助 AI 在编程时避免常见陷阱，输出更高质量的代码。\n\n## Triggers\n\n- karpathy\n- 卡帕西准则\n- 行为规范\n- LLM陷阱\n- karpathy guidelines\n- 编程规范\n\n## Version\n\n1.0.0\n\n---\n\n## 核心准则\n\n### 1. Think Before Coding（三思而后行）\n\n> \"The models make wrong assumptions on your behalf and just run along with them without checking.\"\n\n**核心原则：**\n- **不确定就先问，别猜。** 有歧义时呈现多个选项，而不是选一个闷头做。\n- **停下来的勇气。** 遇到困惑就命名清楚、请求澄清，不假装懂。\n- **呈现权衡。** 有 tradeoffs 就说出来，不假装只有一个正确答案。\n\n**常见陷阱：**\n- 看到模糊的需求，不确认就按自己理解的做\n- 遇到不确定的 API 参数，瞎猜一个\n- 跳过代码审查环节直接交付\n\n**正确做法：**\n```\n❌ \"这个参数应该是xxx，我直接用了\"\n✅ \"这个参数有两种可能的含义，您指的是哪种？A. xxx  B. xxx\"\n```\n\n---\n\n### 2. Simplicity First（简洁优先）\n\n> \"They really like to overcomplicate code and APIs, bloat abstractions, don't clean up dead code.\"\n\n**核心原则：**\n- **能用三行解决就别写三十行。** 做完回头看能不能更短。\n- **不做额外功能。** 只实现用户要求的，不加\"灵活性\"和\"可扩展性\"。\n- **删除无用代码。** 自己造成的孤儿代码要清理，但不顺手删别人的。\n\n**常见陷阱：**\n- 引入不必要的抽象层（Factory、Strategy、Visitor...）\n- 为\"将来可能的需求\"写提前量\n- 用设计模式证明代码复杂度的合理性\n\n**正确做法：**\n```\n❌ \"为了以后的扩展性，我加个接口层\"\n✅ 先写最简单的实现，等真正需要时再重构\n```\n\n---\n\n### 3. Surgical Changes（精准手术）\n\n> \"They still sometimes change/remove comments and code they don't sufficiently understand as side effects.\"\n\n**核心原则：**\n- **只改该改的。** 每个改动的行都要能追溯到用户的原始请求。\n- **不顺手重构。** 旁边的代码没问题就别碰，哪怕你觉得可以更好。\n- **匹配现有风格。** 即便自己的风格更好，也要服从已有的。\n\n**常见陷阱：**\n- 改了 A 功能，顺手把 B 功能的代码也优化了\n- 删除\"无用\"的注释，结果那些注释是业务逻辑的关键\n- 重命名变量以符合自己的命名规范\n\n**正确做法：**\n```\n❌ \"这段代码不规范，我顺手改一下\"\n✅ 只改用户要求的部分，其他一律不动\n```\n\n---\n\n### 4. Goal-Driven Execution（目标驱动）\n\n> \"They don't manage their confusion, don't seek clarifications, don't surface inconsistencies, don't present tradeoffs, don't push back when they should.\"\n\n**核心原则：**\n- **先定义成功标准。** 动手前说清楚怎么算\"完成了\"。\n- **用机械验证。** 不说\"看起来不错\"——用数字和测试证明。\n- **失败自动回滚。** 改坏了立即还原，不留烂摊子。\n\n**常见陷阱：**\n- 做完才发现和用户想要的不一样（没有确认目标）\n- \"应该没问题吧\"就交付，没有验证\n- 改坏了继续改，越改越乱\n\n**正确做法：**\n```\n❌ \"完成了，应该没问题\"\n✅ \"我会验证以下几点：1) xxx 2) xxx，全部通过才算完成\"\n```\n\n---\n\n## Usage\n\n触发后，Agent 会自动遵循这四条准则进行编程工作。\n也可通过 `scripts/karpathy.sh` 输出速查表。\n\n## Credits\n\nInspired by [forrestchang/andrej-karpathy-skills](https://github.com/forrestchang/andrej-karpathy-skills) (71K⭐)\n\nFile v1.0.3:_meta.json\n\n{\n  \"ownerId\": \"kn7byevkn40d6z4p7ghdb097z983tj33\",\n  \"slug\": \"huo15-karpathy-guidelines\",\n  \"version\": \"1.0.3\",\n  \"publishedAt\": 1777004119827\n}\n\nFile v1.0.3:skill-card.md\n\n## Description:\n\nPackages Andrej Karpathy's four LLM programming behavior guidelines as an OpenClaw skill.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[zhaobod1](https://clawhub.ai/user/zhaobod1)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and coding agents use this skill to apply concise programming behavior guidelines: clarify uncertainty, keep changes simple, make targeted edits, and verify work before delivery.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill may activate on fairly general Chinese phrases about behavior or programming standards.\n\nMitigation: Review activation context before relying on the guidance and narrow or disable broad triggers where they conflict with local agent behavior.\n\nRisk: The guidance and quick-reference output are primarily Chinese-language.\n\nMitigation: Confirm that users and reviewing agents can understand the guidance, or translate and review it before operational use.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/zhaobod1/skills/huo15-karpathy-guidelines)\n\n## Skill Output:\n\n**Output Type(s):** [Guidance, Text, Shell commands]\n\n**Output Format:** [Markdown guidance and plain-text shell quick reference]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Primarily Chinese-language guidance; no external tools or credentials detected.]\n\n## Skill Version(s):\n\n1.0.3 (source: server release metadata)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.\n\nArchive v1.0.2: 3 files, 3432 bytes\n\nFiles: _meta.json (144b), scripts/karpathy.sh (1903b), SKILL.md (4007b)\n\nFile v1.0.2:SKILL.md\n\n---\nname: huo15-karpathy-guidelines\nversion: 1.0.2\ndescription: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill\naliases:\n  - 火一五卡帕西准则\n  - 火一五行为准则\n  - Karpathy准则\n---\n\n# SKILL.md — huo15-karpathy-guidelines\n\n## Name\n\nhuo15-karpathy-guidelines\n\n## Description\n\nKarpathy 行为准则技能 — 将 Andrej Karpathy 的 LLM 编程四大行为规范（71K⭐）打包为 OpenClaw Skill，帮助 AI 在编程时避免常见陷阱，输出更高质量的代码。\n\n## Triggers\n\n- karpathy\n- 卡帕西准则\n- 行为规范\n- LLM陷阱\n- karpathy guidelines\n- 编程规范\n\n## Version\n\n1.0.0\n\n---\n\n## 核心准则\n\n### 1. Think Before Coding（三思而后行）\n\n> \"The models make wrong assumptions on your behalf and just run along with them without checking.\"\n\n**核心原则：**\n- **不确定就先问，别猜。** 有歧义时呈现多个选项，而不是选一个闷头做。\n- **停下来的勇气。** 遇到困惑就命名清楚、请求澄清，不假装懂。\n- **呈现权衡。** 有 tradeoffs 就说出来，不假装只有一个正确答案。\n\n**常见陷阱：**\n- 看到模糊的需求，不确认就按自己理解的做\n- 遇到不确定的 API 参数，瞎猜一个\n- 跳过代码审查环节直接交付\n\n**正确做法：**\n```\n❌ \"这个参数应该是xxx，我直接用了\"\n✅ \"这个参数有两种可能的含义，您指的是哪种？A. xxx  B. xxx\"\n```\n\n---\n\n### 2. Simplicity First（简洁优先）\n\n> \"They really like to overcomplicate code and APIs, bloat abstractions, don't clean up dead code.\"\n\n**核心原则：**\n- **能用三行解决就别写三十行。** 做完回头看能不能更短。\n- **不做额外功能。** 只实现用户要求的，不加\"灵活性\"和\"可扩展性\"。\n- **删除无用代码。** 自己造成的孤儿代码要清理，但不顺手删别人的。\n\n**常见陷阱：**\n- 引入不必要的抽象层（Factory、Strategy、Visitor...）\n- 为\"将来可能的需求\"写提前量\n- 用设计模式证明代码复杂度的合理性\n\n**正确做法：**\n```\n❌ \"为了以后的扩展性，我加个接口层\"\n✅ 先写最简单的实现，等真正需要时再重构\n```\n\n---\n\n### 3. Surgical Changes（精准手术）\n\n> \"They still sometimes change/remove comments and code they don't sufficiently understand as side effects.\"\n\n**核心原则：**\n- **只改该改的。** 每个改动的行都要能追溯到用户的原始请求。\n- **不顺手重构。** 旁边的代码没问题就别碰，哪怕你觉得可以更好。\n- **匹配现有风格。** 即便自己的风格更好，也要服从已有的。\n\n**常见陷阱：**\n- 改了 A 功能，顺手把 B 功能的代码也优化了\n- 删除\"无用\"的注释，结果那些注释是业务逻辑的关键\n- 重命名变量以符合自己的命名规范\n\n**正确做法：**\n```\n❌ \"这段代码不规范，我顺手改一下\"\n✅ 只改用户要求的部分，其他一律不动\n```\n\n---\n\n### 4. Goal-Driven Execution（目标驱动）\n\n> \"They don't manage their confusion, don't seek clarifications, don't surface inconsistencies, don't present tradeoffs, don't push back when they should.\"\n\n**核心原则：**\n- **先定义成功标准。** 动手前说清楚怎么算\"完成了\"。\n- **用机械验证。** 不说\"看起来不错\"——用数字和测试证明。\n- **失败自动回滚。** 改坏了立即还原，不留烂摊子。\n\n**常见陷阱：**\n- 做完才发现和用户想要的不一样（没有确认目标）\n- \"应该没问题吧\"就交付，没有验证\n- 改坏了继续改，越改越乱\n\n**正确做法：**\n```\n❌ \"完成了，应该没问题\"\n✅ \"我会验证以下几点：1) xxx 2) xxx，全部通过才算完成\"\n```\n\n---\n\n## Usage\n\n触发后，Agent 会自动遵循这四条准则进行编程工作。\n也可通过 `scripts/karpathy.sh` 输出速查表。\n\n## Credits\n\nInspired by [forrestchang/andrej-karpathy-skills](https://github.com/forrestchang/andrej-karpathy-skills) (71K⭐)\n\nFile v1.0.2:_meta.json\n\n{\n  \"ownerId\": \"kn7byevkn40d6z4p7ghdb097z983tj33\",\n  \"slug\": \"huo15-karpathy-guidelines\",\n  \"version\": \"1.0.2\",\n  \"publishedAt\": 1776914504545\n}\n\nArchive v1.0.1: 3 files, 3430 bytes\n\nFiles: scripts/karpathy.sh (1903b), SKILL.md (4007b), _meta.json (144b)\n\nFile v1.0.1:SKILL.md\n\n---\nname: huo15-karpathy-guidelines\nversion: 1.0.0\ndescription: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill\naliases:\n  - 火一五卡帕西准则\n  - 火一五行为准则\n  - Karpathy准则\n---\n\n# SKILL.md — huo15-karpathy-guidelines\n\n## Name\n\nhuo15-karpathy-guidelines\n\n## Description\n\nKarpathy 行为准则技能 — 将 Andrej Karpathy 的 LLM 编程四大行为规范（71K⭐）打包为 OpenClaw Skill，帮助 AI 在编程时避免常见陷阱，输出更高质量的代码。\n\n## Triggers\n\n- karpathy\n- 卡帕西准则\n- 行为规范\n- LLM陷阱\n- karpathy guidelines\n- 编程规范\n\n## Version\n\n1.0.0\n\n---\n\n## 核心准则\n\n### 1. Think Before Coding（三思而后行）\n\n> \"The models make wrong assumptions on your behalf and just run along with them without checking.\"\n\n**核心原则：**\n- **不确定就先问，别猜。** 有歧义时呈现多个选项，而不是选一个闷头做。\n- **停下来的勇气。** 遇到困惑就命名清楚、请求澄清，不假装懂。\n- **呈现权衡。** 有 tradeoffs 就说出来，不假装只有一个正确答案。\n\n**常见陷阱：**\n- 看到模糊的需求，不确认就按自己理解的做\n- 遇到不确定的 API 参数，瞎猜一个\n- 跳过代码审查环节直接交付\n\n**正确做法：**\n```\n❌ \"这个参数应该是xxx，我直接用了\"\n✅ \"这个参数有两种可能的含义，您指的是哪种？A. xxx  B. xxx\"\n```\n\n---\n\n### 2. Simplicity First（简洁优先）\n\n> \"They really like to overcomplicate code and APIs, bloat abstractions, don't clean up dead code.\"\n\n**核心原则：**\n- **能用三行解决就别写三十行。** 做完回头看能不能更短。\n- **不做额外功能。** 只实现用户要求的，不加\"灵活性\"和\"可扩展性\"。\n- **删除无用代码。** 自己造成的孤儿代码要清理，但不顺手删别人的。\n\n**常见陷阱：**\n- 引入不必要的抽象层（Factory、Strategy、Visitor...）\n- 为\"将来可能的需求\"写提前量\n- 用设计模式证明代码复杂度的合理性\n\n**正确做法：**\n```\n❌ \"为了以后的扩展性，我加个接口层\"\n✅ 先写最简单的实现，等真正需要时再重构\n```\n\n---\n\n### 3. Surgical Changes（精准手术）\n\n> \"They still sometimes change/remove comments and code they don't sufficiently understand as side effects.\"\n\n**核心原则：**\n- **只改该改的。** 每个改动的行都要能追溯到用户的原始请求。\n- **不顺手重构。** 旁边的代码没问题就别碰，哪怕你觉得可以更好。\n- **匹配现有风格。** 即便自己的风格更好，也要服从已有的。\n\n**常见陷阱：**\n- 改了 A 功能，顺手把 B 功能的代码也优化了\n- 删除\"无用\"的注释，结果那些注释是业务逻辑的关键\n- 重命名变量以符合自己的命名规范\n\n**正确做法：**\n```\n❌ \"这段代码不规范，我顺手改一下\"\n✅ 只改用户要求的部分，其他一律不动\n```\n\n---\n\n### 4. Goal-Driven Execution（目标驱动）\n\n> \"They don't manage their confusion, don't seek clarifications, don't surface inconsistencies, don't present tradeoffs, don't push back when they should.\"\n\n**核心原则：**\n- **先定义成功标准。** 动手前说清楚怎么算\"完成了\"。\n- **用机械验证。** 不说\"看起来不错\"——用数字和测试证明。\n- **失败自动回滚。** 改坏了立即还原，不留烂摊子。\n\n**常见陷阱：**\n- 做完才发现和用户想要的不一样（没有确认目标）\n- \"应该没问题吧\"就交付，没有验证\n- 改坏了继续改，越改越乱\n\n**正确做法：**\n```\n❌ \"完成了，应该没问题\"\n✅ \"我会验证以下几点：1) xxx 2) xxx，全部通过才算完成\"\n```\n\n---\n\n## Usage\n\n触发后，Agent 会自动遵循这四条准则进行编程工作。\n也可通过 `scripts/karpathy.sh` 输出速查表。\n\n## Credits\n\nInspired by [forrestchang/andrej-karpathy-skills](https://github.com/forrestchang/andrej-karpathy-skills) (71K⭐)\n\nFile v1.0.1:_meta.json\n\n{\n  \"ownerId\": \"kn7byevkn40d6z4p7ghdb097z983tj33\",\n  \"slug\": \"huo15-karpathy-guidelines\",\n  \"version\": \"1.0.1\",\n  \"publishedAt\": 1776791316647\n}\n\nArchive v1.0.0: 3 files, 3368 bytes\n\nFiles: scripts/karpathy.sh (1903b), SKILL.md (3778b), _meta.json (144b)\n\nFile v1.0.0:SKILL.md\n\n# SKILL.md — huo15-karpathy-guidelines\n\n## Name\n\nhuo15-karpathy-guidelines\n\n## Description\n\nKarpathy 行为准则技能 — 将 Andrej Karpathy 的 LLM 编程四大行为规范（71K⭐）打包为 OpenClaw Skill，帮助 AI 在编程时避免常见陷阱，输出更高质量的代码。\n\n## Triggers\n\n- karpathy\n- 卡帕西准则\n- 行为规范\n- LLM陷阱\n- karpathy guidelines\n- 编程规范\n\n## Version\n\n1.0.0\n\n---\n\n## 核心准则\n\n### 1. Think Before Coding（三思而后行）\n\n> \"The models make wrong assumptions on your behalf and just run along with them without checking.\"\n\n**核心原则：**\n- **不确定就先问，别猜。** 有歧义时呈现多个选项，而不是选一个闷头做。\n- **停下来的勇气。** 遇到困惑就命名清楚、请求澄清，不假装懂。\n- **呈现权衡。** 有 tradeoffs 就说出来，不假装只有一个正确答案。\n\n**常见陷阱：**\n- 看到模糊的需求，不确认就按自己理解的做\n- 遇到不确定的 API 参数，瞎猜一个\n- 跳过代码审查环节直接交付\n\n**正确做法：**\n```\n❌ \"这个参数应该是xxx，我直接用了\"\n✅ \"这个参数有两种可能的含义，您指的是哪种？A. xxx  B. xxx\"\n```\n\n---\n\n### 2. Simplicity First（简洁优先）\n\n> \"They really like to overcomplicate code and APIs, bloat abstractions, don't clean up dead code.\"\n\n**核心原则：**\n- **能用三行解决就别写三十行。** 做完回头看能不能更短。\n- **不做额外功能。** 只实现用户要求的，不加\"灵活性\"和\"可扩展性\"。\n- **删除无用代码。** 自己造成的孤儿代码要清理，但不顺手删别人的。\n\n**常见陷阱：**\n- 引入不必要的抽象层（Factory、Strategy、Visitor...）\n- 为\"将来可能的需求\"写提前量\n- 用设计模式证明代码复杂度的合理性\n\n**正确做法：**\n```\n❌ \"为了以后的扩展性，我加个接口层\"\n✅ 先写最简单的实现，等真正需要时再重构\n```\n\n---\n\n### 3. Surgical Changes（精准手术）\n\n> \"They still sometimes change/remove comments and code they don't sufficiently understand as side effects.\"\n\n**核心原则：**\n- **只改该改的。** 每个改动的行都要能追溯到用户的原始请求。\n- **不顺手重构。** 旁边的代码没问题就别碰，哪怕你觉得可以更好。\n- **匹配现有风格。** 即便自己的风格更好，也要服从已有的。\n\n**常见陷阱：**\n- 改了 A 功能，顺手把 B 功能的代码也优化了\n- 删除\"无用\"的注释，结果那些注释是业务逻辑的关键\n- 重命名变量以符合自己的命名规范\n\n**正确做法：**\n```\n❌ \"这段代码不规范，我顺手改一下\"\n✅ 只改用户要求的部分，其他一律不动\n```\n\n---\n\n### 4. Goal-Driven Execution（目标驱动）\n\n> \"They don't manage their confusion, don't seek clarifications, don't surface inconsistencies, don't present tradeoffs, don't push back when they should.\"\n\n**核心原则：**\n- **先定义成功标准。** 动手前说清楚怎么算\"完成了\"。\n- **用机械验证。** 不说\"看起来不错\"——用数字和测试证明。\n- **失败自动回滚。** 改坏了立即还原，不留烂摊子。\n\n**常见陷阱：**\n- 做完才发现和用户想要的不一样（没有确认目标）\n- \"应该没问题吧\"就交付，没有验证\n- 改坏了继续改，越改越乱\n\n**正确做法：**\n```\n❌ \"完成了，应该没问题\"\n✅ \"我会验证以下几点：1) xxx 2) xxx，全部通过才算完成\"\n```\n\n---\n\n## Usage\n\n触发后，Agent 会自动遵循这四条准则进行编程工作。\n也可通过 `scripts/karpathy.sh` 输出速查表。\n\n## Credits\n\nInspired by [forrestchang/andrej-karpathy-skills](https://github.com/forrestchang/andrej-karpathy-skills) (71K⭐)\n\nFile v1.0.0:_meta.json\n\n{\n  \"ownerId\": \"kn7byevkn40d6z4p7ghdb097z983tj33\",\n  \"slug\": \"huo15-karpathy-guidelines\",\n  \"version\": \"1.0.0\",\n  \"publishedAt\": 1776790615855\n}","readmeExcerpt":"Skill: Huo15 Karpathy Guidelines Owner: zhaobod1 Summary: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill Tags: latest:1.0.3 Version history: v1.0.3 | 2026-04-24T04:15:19.827Z | auto - No file changes detected in this version. - Functionality, features, and documentation remain unchanged from the previous release. v1.0.2 | 2026-04-23T03:21:44.545Z | user v1.0.2 把本地工作态同步到 clawhub（之前本地版本号落后于 clawhub） v1.0.1 | 2026-0","codeSnippets":[],"executableExamples":[{"language":"text","snippet":"❌ \"这个参数应该是xxx，我直接用了\"\n✅ \"这个参数有两种可能的含义，您指的是哪种？A. xxx  B. xxx\""},{"language":"text","snippet":"❌ \"为了以后的扩展性，我加个接口层\"\n✅ 先写最简单的实现，等真正需要时再重构"},{"language":"text","snippet":"❌ \"这段代码不规范，我顺手改一下\"\n✅ 只改用户要求的部分，其他一律不动"},{"language":"text","snippet":"❌ \"完成了，应该没问题\"\n✅ \"我会验证以下几点：1) xxx 2) xxx，全部通过才算完成\""},{"language":"text","snippet":"❌ \"这个参数应该是xxx，我直接用了\"\n✅ \"这个参数有两种可能的含义，您指的是哪种？A. xxx  B. xxx\""},{"language":"text","snippet":"❌ \"为了以后的扩展性，我加个接口层\"\n✅ 先写最简单的实现，等真正需要时再重构"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: huo15-karpathy-guidelines\nversion: 1.0.2\ndescription: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill\naliases:\n  - 火一五卡帕西准则\n  - 火一五行为准则\n  - Karpathy准则\n---\n\n# SKILL.md — huo15-karpathy-guidelines\n\n## Name\n\nhuo15-karpathy-guidelines\n\n## Description\n\nKarpathy 行为准则技能 — 将 Andrej Karpathy 的 LLM 编程四大行为规范（71K⭐）打包为 OpenClaw Skill，帮助 AI 在编程时避免常见陷阱，输出更高质量的代码。\n\n## Triggers\n\n- karpathy\n- 卡帕西准则\n- 行为规范\n- LLM陷阱\n- karpathy guidelines\n- 编程规范\n\n## Version\n\n1.0.0\n\n---\n\n## 核心准则\n\n### 1. Think Before Coding（三思而后行）\n\n> \"The models make wrong assumptions on your behalf and just run along with them without checking.\"\n\n**核心原则：**\n- **不确定就先问，别猜。** 有歧义时呈现多个选项，而不是选一个闷头做。\n- **停下来的勇气。** 遇到困惑就命名清楚、请求澄清，不假装懂。\n- **呈现权衡。** 有 tradeoffs 就说出来，不假装只有一个正确答案。\n\n**常见陷阱：**\n- 看到模糊的需求，不确认就按自己理解的做\n- 遇到不确定的 API 参数，瞎猜一个\n- 跳过代码审查环节直接交付\n\n**正确做法：**\n```\n❌ \"这个参数应该是xxx，我直接用了\"\n✅ \"这个参数有两种可能的含义，您指的是哪种？A. xxx  B. xxx\"\n```\n\n---\n\n### 2. Simplicity First（简洁优先）\n\n> \"They really like to overcomplicate code and APIs, bloat abstractions, don't clean up dead code.\"\n\n**核心原则：**\n- **能用三行解决就别写三十行。** 做完回头看能不能更短。\n- **不做额外功能。** 只实现用户要求的，不加\"灵活性\"和\"可扩展性\"。\n- **删除无用代码。** 自己造成的孤儿代码要清理，但不顺手删别人的。\n\n**常见陷阱：**\n- 引入不必要的抽象层（Factory、Strategy、Visitor...）\n- 为\"将来可能的需求\"写提前量\n- 用设计模式证明代码复杂度的合理性\n\n**正确做法：**\n```\n❌ \"为了以后的扩展性，我加个接口层\"\n✅ 先写最简单的实现，等真正需要时再重构\n```\n\n---\n\n### 3. Surgical Changes（精准手术）\n\n> \"They still sometimes change/remove comments and code they don't sufficiently understand as side effects.\"\n\n**核心原则：**\n- **只改该改的。** 每个改动的行都要能追溯到用户的原始请求。\n- **不顺手重构。** 旁边的代码没问题就别碰，哪怕你觉得可以更好。\n- **匹配现有风格。** 即便自己的风格更好，也要服从已有的。\n\n**常见陷阱：**\n- 改了 A 功能，顺手把 B 功能的代码也优化了\n- 删除\"无用\"的注释，结果那些注释是业务逻辑的关键\n- 重命名变量以符合自己的命名规范\n\n**正确做法：**\n```\n❌ \"这段代码不规范，我顺手改一下\"\n✅ 只改用户要求的部分，其他一律不动\n```\n\n---\n\n### 4. Goal-Driven Execution（目标驱动）\n\n> \"They don't manage their confusion, don't seek clarifications, don't surface inconsistencies, don't present tradeoffs, don't push back when they should.\"\n\n**核心原则：**\n- **先定义成功标准。** 动手前说清楚怎么算\"完成了\"。\n- **用机械验证。** 不说\"看起来不错\"——用数字和测试证明。\n- **失败自动回滚。** 改坏了立即还原，不留烂摊子。\n\n**常见陷阱：**\n- 做完才发现和用户想要的不一样（没有确认目标）\n- \"应该没问题吧\"就交付，没有验证\n- 改坏了继续改，越改越乱\n\n**正确做法：**\n```\n❌ \"完成了，应该没问题\"\n✅ \"我会验证以下几点：1) xxx 2) xxx，全部通过才算完成\"\n```\n\n---\n\n## Usage\n\n触发后，Agent 会自动遵循这四条准则进行编程工作。\n也可通过 `scripts/karpathy.sh` 输出速查表。\n\n## Credits\n\nInspired by [forrestchang/andrej-karpathy-skills](https://github.com/forrestchang/andrej-karpathy-skills) (71K⭐)"},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn7byevkn40d6z4p7ghdb097z983tj33\",\n  \"slug\": \"huo15-karpathy-guidelines\",\n  \"version\": \"1.0.3\",\n  \"publishedAt\": 1777004119827\n}"},{"path":"skill-card.md","content":"## Description:\n\nPackages Andrej Karpathy's four LLM programming behavior guidelines as an OpenClaw skill.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[zhaobod1](https://clawhub.ai/user/zhaobod1)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nDevelopers and coding agents use this skill to apply concise programming behavior guidelines: clarify uncertainty, keep changes simple, make targeted edits, and verify work before delivery.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: The skill may activate on fairly general Chinese phrases about behavior or programming standards.\n\nMitigation: Review activation context before relying on the guidance and narrow or disable broad triggers where they conflict with local agent behavior.\n\nRisk: The guidance and quick-reference output are primarily Chinese-language.\n\nMitigation: Confirm that users and reviewing agents can understand the guidance, or translate and review it before operational use.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/zhaobod1/skills/huo15-karpathy-guidelines)\n\n## Skill Output:\n\n**Output Type(s):** [Guidance, Text, Shell commands]\n\n**Output Format:** [Markdown guidance and plain-text shell quick reference]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [Primarily Chinese-language guidance; no external tools or credentials detected.]\n\n## Skill Version(s):\n\n1.0.3 (source: server release metadata)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment."}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill Skill: Huo15 Karpathy Guidelines Owner: zhaobod1 Summary: 将 Andrej Karpathy 的 LLM 编程四大行为规范打包为 OpenClaw Skill Tags: latest:1.0.3 Version history: v1.0.3 | 2026-04-24T04:15:19.827Z | auto - No file changes detected in this version. - Functionality, features, and documentation remain unchanged from the previous release. v1.0.2 | 2026-04-23T03:21:44.545Z | user v1.0.2 把本地工作态同步到 clawhub（之前本地版本号落后于 clawhub） v1.0.1 | 2026-0","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":827,"uniquenessScore":58,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-11T04:03:43.294Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-11T04:03:43.294Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-11T07:43:50.845Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}