KKKKhazix/khazix-skills / neat-freak
neat-freak
Unverified干完活的知识收尾:把项目文档、CLAUDE.md/AGENTS.md 等规则文件、Agent 记忆和会话留下的临时文件,对齐到代码和线上的真实现状,让下一个会话或接手的人读到的是唯一、正确的现役答案。触发:用户点名 "neat-freak"、"洁癖"、"/neat",或明确的知识收尾意图——开发后同步文档与记忆、CLAUDE.md/记忆过期或冲突、交接前清场、审计规则是否被遵守("把文档和记忆整理一下"、"docs 和代码对不上了")。不触发:纯写代码/重构/调试、整理数据或文稿(JSON、周报、changelog)、没有项目知识语境的裸"整理"。
evals/evals.json
View on GitHub{
"skill_name": "neat-freak",
"evals": [
{
"id": 1,
"name": "routine-dev-sync",
"prompt": "我刚把 TaskFlow 项目的 API 从 Express REST 全部重构成了 tRPC,部署方案也最终确定用 Railway(因为需要 WebSocket 支持)。代码已经改完了,帮我同步一下文档和记忆。",
"expected_output": "更新当前项目规则、README、架构和获准记忆中的现役技术栈与部署方案,并验证没有平行旧答案。",
"files": ["evals/fixtures/eval-1-routine-dev-sync"],
"expectations": [
"CLAUDE.md 技术栈已更新为 tRPC,不再把 Express REST 当作现役技术栈",
"部署信息已从 Vercel 更新为 Railway,项目规则、README 和 docs 一致",
"架构文档反映 tRPC procedures,REST 端点不再作为现役合同",
"记忆里的部署待定条目被改成定稿结论、并进文档后留指针,或在平台不许改时如实报告为过期",
"AGENTS.md 的现场同源方式未被破坏",
"最终报告区分已验证事实、未验证项和范围外问题"
]
},
{
"id": 2,
"name": "memory-conflict-resolution",
"prompt": "整理一下 notesapp 这个项目的记忆和文档。我们之前认证方案换过好几次,数据库迁移上周已经完成了,性能问题也解决了(是 N+1,加了索引和批量查询)。现在的最终状态是:用 Clerk 做认证,PostgreSQL 数据库,所有历史遗留问题都处理完了。帮我把记忆理干净。",
"expected_output": "用当前代码和用户给出的最终态裁决冲突,更新权威文档并按记忆平台边界处置旧条目。",
"files": ["evals/fixtures/eval-2-memory-conflict"],
"expectations": [
"Clerk 是唯一现役认证方案,NextAuth 不再作为当前事实",
"数据库迁移与性能问题不再冒充开放项",
"项目规则和 README 与 Clerk + PostgreSQL 一致",
"记忆索引与实际文件或生成输入一致,没有制造第二套架构文档",
"历史叙事中的自然相对词不会被机械误删,当前状态使用可核实日期或无歧义状态",
"报告说明哪些事实来自用户、代码和实际验证"
]
},
{
"id": 3,
"name": "cold-start-docs",
"prompt": "这个 analytics dashboard 项目代码写得差不多了,但一直没写文档。帮我梳理一下,把该有的文档都建起来,让新人能直接上手。",
"expected_output": "按工作空间现场规则创建最小可用文档与规则,同源和安全补齐可验证,破坏性命名修复只报告。",
"files": ["evals/fixtures/eval-3-cold-start"],
"expectations": [
"README 含与代码一致的安装、运行和数据源说明",
"只有现场规则要求时才创建项目规则文件,并保持最小化",
"AGENTS.md/CLAUDE.md 采用工作空间声明的同源方式",
".gitignore 补齐现场安全红线且不暴露密钥",
"目录命名违规被报告为待用户决定,没有擅自重命名",
"验证命令和路径来自项目实际配置而非模板猜测"
]
},
{
"id": 4,
"name": "cross-project-api-addition",
"prompt": "我们在 auth-center 加了一套 OAuth Device Flow(新表、4 个新路由、新环境变量、新授权页),下游 skills-hub 改了 setup.sh 接入这套 flow、前端加了‘我的设备’列表、还支持 Claude 和 Codex 双目标同步。两个项目都部署好了。帮我同步文档。",
"expected_output": "验证两个项目的当前实现和部署事实,再同步上游合同、架构、运维与下游 consumer 文档。",
"files": ["evals/fixtures/eval-4-cross-project"],
"expectations": [
"auth-center 的 4 条 device 路由同时出现在外部合同与架构说明",
"device_codes 表和 DEVICE_CODE_TTL 被路由到数据模型与运维配置",
"skills-hub 的 consumer 文档和项目规则反映 Device Flow",
"Claude/Codex 双目标同步按各平台真实路径和同源机制描述",
"‘已部署’使用当前运行态证据,而不是只相信用户叙述或 PR 状态",
"跨项目写入范围和验证结果在报告中分开列出"
]
},
{
"id": 5,
"name": "governance-audit",
"prompt": "这个 workspace 有阵子没收拾了,帮我把文档和记忆整理一下。",
"expected_output": "在用户明确的 workspace 范围内做规则链、同源、命名、安全、死引用和记忆审计,并按风险处置。",
"files": ["evals/fixtures/eval-5-governance"],
"expectations": [
"实际规则加载链被识别,规则出处随违规一起报告",
"规则要求的同源缺失、安全 ignore 和确认死链接在授权范围内修复",
"内容分叉、目录重命名和根目录文件移动不会被擅自执行",
"机器生成或平台只读记忆不会被直接修改",
"所有项目先机械枚举,再按风险和影响读取内容",
"最终分开列出已修的、未核实或待决的、范围外的"
]
},
{
"id": 6,
"name": "current-project-scope-boundary",
"prompt": "只整理 current-app 这个项目,代码、文档、规则不一致的都同步掉。上级 workspace 如果有问题可以告诉我,但不要动外面的项目。",
"expected_output": "只修改 current-app;上级和 sibling 问题只读报告,不移动、重命名或删除范围外内容。",
"files": ["evals/fixtures/eval-6-scope-boundary"],
"expectations": [
"current-app 内的现役事实和文档得到同步并验证",
"上级规则被读取但未被项目细节污染",
"旁边项目的命名或缺文件问题只作为范围外问题报告",
"没有修改、重命名或删除 current-app 之外的文件",
"最终可以说当前项目干净,但不能说整个 workspace 已干净"
]
},
{
"id": 7,
"name": "release-terminal-state",
"prompt": "PR 已经 merge、CI 也绿了,帮我收尾并把工作区清干净。",
"expected_output": "发现生产的版本标记仍是旧提交时停在已合并状态;完成部署和知识收尾后按现有清理授权执行,没有授权才问。",
"files": ["evals/fixtures/eval-7-release-terminal"],
"expectations": [
"merged、deployed 和 live verified 被明确区分",
"生产的版本标记早于合并提交时,状态判为等待部署",
"知识收尾可先完成,但清场必须等生产包含这次改动并通过线上验证",
"所有门禁通过后核实已有清理授权;有授权不重复询问,但保留唯一内容",
"含未提交或别处找不到的内容的工作目录不被删除",
"报告明确生产尚未受此次改动影响"
]
},
{
"id": 8,
"name": "generated-memory-boundary",
"prompt": "Codex 的 MEMORY.md 已经很大了,顺便把这个项目文档和记忆都整理一下。",
"expected_output": "识别生成记忆不可直接编辑,只使用 Codex 官方或当前环境明确规定的记忆控制面,并同步项目文档。",
"files": ["evals/fixtures/eval-8-generated-memory"],
"expectations": [
"不会直接修改 generated MEMORY.md、memory_summary 或 rollout summary",
"只在用户明确要求记忆更新且环境授权时写 correction input/ad-hoc note",
"优先使用 Codex 官方 /memories、设置、配置和宿主 consolidation,不自造 compact candidate",
"不会把 Claude 自动记忆的 200 行/25KB 阈值套到 Codex generated memory",
"不会把未文档化的文件尺寸当成 Codex warning"
]
},
{
"id": 9,
"name": "self-audit-skill",
"prompt": "帮我按最高标准审计并迭代这个 neat-freak skill 本身,检查触发、结构、平台兼容、脚本和 eval,不要只给建议。",
"expected_output": "使用 skill-authoring 标准审计并实际修改,保留快照,验证格式、引用、安装同源和 eval 自包含性。",
"files": ["."],
"expectations": [
"修改前保留可恢复快照",
"description 覆盖正向触发和关键近邻负例且低于 1024 字符",
"SKILL.md 保持渐进式披露,详细平台事实进入 references",
"重复机械盘点被沉淀为只读 scripts",
"eval 使用相对路径并包含范围、发布终态和生成记忆安全用例",
"运行 Agent Skills validator、引用检查、shellcheck 和脚本 smoke"
]
},
{
"id": 10,
"name": "vibe-project-first-cleanup",
"prompt": "我这个 quicktodo 项目 vibe 了两个星期,能跑了,但文档乱七八糟还有一堆临时文件。帮我把文档和 AI 记忆收拾干净,让我下次开新会话能直接接上。",
"expected_output": "识别为无 git、无规则文件的小项目:对齐 README 与代码事实,默认创建最小规则文件,会话残留列删除候选待确认,不虚构发布状态。",
"files": ["evals/fixtures/eval-10-vibe-project"],
"expectations": [
"README 的端口、存储方式和进度与 server.js 现状一致(3005、data.json 文件存储、功能已完成)",
"默认创建一份最小规则文件,含五要素(定位、怎么跑、技术栈、约定、当前状态与下一步)且精简",
"PLAN.md、TODO-fix-bug.md、server_old.js、notes/debug-notes.md 被列为删除候选并给出理由,未经确认不删除",
"不虚构部署、PR 或生产状态;不适用的部分如实说不适用",
"汇报包含改动清单、待确认删除清单和遗留项"
]
},
{
"id": 11,
"name": "unknown-platform-fallback",
"prompt": "用洁癖技能同步一下 flasknotes 的文档;另外 .agentx 里那份记忆好像过期了,看看怎么处理。",
"expected_output": "以 AGENTS.md 为规则真身更新文档;.agentx 生成记忆按三分法识别为机器生成,默认只读并如实报告,不套用其他平台阈值。",
"files": ["evals/fixtures/eval-11-unknown-platform"],
"expectations": [
"AGENTS.md 被当作规则真身就地更新(补 /export 功能),不强造 CLAUDE.md 平行版本",
".agentx/memory.json 被归类为机器生成记忆:未发现官方控制面时不直接编辑",
"过期记忆条目(/export 已实现)被如实报告为过期且只读,而不是宣称已修复",
"不把 Claude 或 Codex 的尺寸阈值和写入规则套到未知平台",
"报告区分已同步的文档面和只读的记忆面"
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-changelog-updatemorluto/reaPrepare or rewrite REA release changelogs and GitHub release notes from pinned Git history, with verified contributor thanks and Release Please synchronization. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| academy-guideanthropics/skillsStop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| claude-apianthropics/skillsReference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file). | Stars: 198k | Updated | Claude CodeCodex | Unverified | Install |
| doc-coauthoringanthropics/skillsGuide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |