[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M71JH0RSW56P8ACH9G3B4922":3,"skill-same:01M71JH0RSW56P8ACH9G3B4922":97},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":16,"sourceHtmlUrl":17,"categories":18,"keywords":19,"publishedAt":39,"linkOnly":40,"markdownContent":41,"markdownBody":42,"frontMatter":43,"refName":44,"commitSha":45,"skillDirectory":5,"forksCount":46,"repositoryHtmlUrl":47,"repositoryHomepageUrl":48,"ownerHtmlUrl":49,"ownerBlogUrl":50,"ownerDescription":51,"updatedAt":52,"install":53,"resources":66},"01M71JH0RSW56P8ACH9G3B4922","how",null,"Use for \"how does X work\", code walkthroughs before changing something, and placement \u002F ownership \u002F layering questions (\"where should this live\", \"which package owns this\", \"is this the right layer\"). Explains subsystem architecture, runtime flow, onboarding mental models. Can critique architecture. Use why for motivation.","caezium\u002Fskills","caezium","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F113233555?v=4","01M71JB77ZPBAWEK009Y2W8BP7","Henry Zhang","01M71JB792NVXCWSYD27HQ0964",3,"MIT","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002FSKILL.md",[],[5,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38],"work","onboarding","ownership","live","layer","flow","questions","why","models","critique","can","architecture","code","use","package","runtime","for","layering","which","2026-11-10 10:12:30",false,"---\nname: how\ndescription: \"Use for \\\"how does X work\\\", code walkthroughs before changing something, and placement \u002F ownership \u002F layering questions (\\\"where should this live\\\", \\\"which package owns this\\\", \\\"is this the right layer\\\"). Explains subsystem architecture, runtime flow, onboarding mental models. Can critique architecture. Use why for motivation.\"\n---\n\n# How\n\nExplore the codebase to answer \"how does X work?\" questions. Produce clear architectural explanations at the level of a senior engineer onboarding onto a subsystem. Enough to build a working mental model, not annotated source code.\n\nTwo modes:\n\n1. **Explain** (default). Explore the codebase and produce a clear explanation\n2. **Critique.** Explain first, then run independent architectural reviews when the host permits them\n\n## Explain Mode\n\n### Step 1. Understand the Question and Assess Complexity\n\nParse what the user is asking about:\n\n- \"How does the rate limiter work?\", a subsystem\n- \"How do we handle billing for on-demand usage?\", a feature flow\n- \"How is the auth service structured?\", an architectural overview\n- \"Walk me through what happens when a user submits a form\", a runtime trace\n\nIdentify the scope. If ambiguous, state your best-guess interpretation before exploring. Don't ask. Let the user redirect if you're off.\n\n**Assess complexity to decide the approach:**\n\n- **Simple** (a single module, a small utility, a narrow question like \"how does function X work\"): skip explorer agents; the explainer explores and explains in a single pass. Go to Step 2b.\n- **Complex** (a subsystem spanning multiple files\u002Fservices, a cross-cutting feature, a full architectural overview): use parallel explorer agents when the user and host policy permit delegation; otherwise explore the same slices inline. Go to Step 2a.\n\nWhen in doubt, lean simple. You can widen the exploration if the first pass hits a wall.\n\n### Step 2a. Explore (complex questions only)\n\nDecompose the question into 2-4 parallel exploration angles, each a distinct slice of the subsystem so explorers don't duplicate work. Example split for \"how does the rate limiter work?\":\n\n- Explorer 1: data model and state management\n- Explorer 2: request path and enforcement\n- Explorer 3: configuration and metrics infrastructure\n\nThe right decomposition depends on the question. Use your judgment. Narrow questions: 2 explorers is fine. Broad subsystems: up to 4.\n\nWhen delegation is authorized, launch the explorers together with non-overlapping briefs, the inherited model, and read-only posture. Do not request a named model or broader permissions. When delegation is unavailable, work through the same exploration angles inline and keep their findings separate until synthesis.\n\nEach explorer gets the same base prompt from `references\u002Fexplorer-prompt.md` plus a specific exploration angle naming its slice. Each explorer should:\n- Start broad: Glob for relevant directories, Grep for key types\u002Finterfaces\u002Fclass names\n- Follow the thread: from an entry point, trace the call chain (callers, callees, data flow, type definitions)\n- Read the actual code, don't guess from file names\n- Stop when it can describe the full path from input to output (or trigger to effect) without hand-waving any step\n- Note things that are surprising, non-obvious, or that a newcomer would get wrong\n\nEach explorer returns structured findings: components found, flow traced, files read, anything non-obvious. Overlap between explorers is fine; the explainer reconciles.\n\nThen proceed to Step 3.\n\n### Step 2b. Direct Explain (simple questions)\n\nExplore and explain in one pass. Read `references\u002Fexplainer-prompt.md` for the communication style and output format. Use repository search and file reads directly; there is no need to delegate a narrow question.\n\nProceed to Step 4.\n\n### Step 3. Synthesize (complex questions only)\n\nSynthesize the exploration slices into one coherent explanation. You may use one independent explainer when delegation is already authorized and the subsystem is large; otherwise synthesize inline. Read `references\u002Fexplainer-prompt.md` for the full prompt template. Reconcile overlap and contradictions before presenting the result.\n\n### Step 4. Present\n\nPresent the explainer's output to the user. You may lightly edit for clarity or add context from the conversation, but don't substantially rewrite. The explainer's communication is the product.\n\n### Output Format\n\nFollow this structure, adapted to the question. Not every section is needed for every question.\n\n**Overview.** 1-2 paragraphs. What it is, what it does, why it exists. Enough to decide whether to keep reading.\n\n**Key Concepts.** The important types, services, or abstractions. Brief definition of each. Not exhaustive, just the ones needed to understand the rest.\n\n**How It Works.** The core of the explanation. Walk through the flow: what triggers it, what happens step by step, where data goes, the decision points. Prose, not pseudocode. Reference specific files and functions so the reader can go look, but don't dump code blocks unless a snippet is genuinely necessary.\n\n**Where Things Live.** A brief map of the relevant files\u002Fdirectories. Not every file, just the ones needed to start working in this area.\n\n**Gotchas.** Non-obvious or surprising things that would trip someone up. Historical context that explains why something looks weird. Known sharp edges.\n\n## Critique Mode\n\nTriggered when the user asks for architectural issues, problems, or improvements, not just understanding.\n\n### Step 1. Explain First\n\nRun the full explain flow above (Steps 1-4). You must understand the architecture before critiquing it.\n\n### Step 2. Run independent critiques\n\nAfter the explanation is complete, run independent architectural critiques. Use parallel critics only when the user and host policy permit delegation, and use the host's inherited model unless the user explicitly asks for model diversity. Otherwise perform one fresh critique pass after setting the explanation aside.\n\nRead `references\u002Fcritic-prompt.md` for the prompt template. Each critic gets:\n1. The explanation from Step 1 (so they don't re-explore)\n2. The relevant file paths (so they can read the actual code)\n3. The architectural critique rubric from `references\u002Fcritique-rubric.md`\n\n### Step 3. Lead Judgment\n\nSame framework as the interrogate skill. You're a pragmatic lead, not an aggregator.\n\nCategorize findings:\n- **Act on.** Architectural problems worth fixing now\n- **Consider.** Real concerns, but the cost\u002Fbenefit is unclear\n- **Noted.** Valid observations, low priority\n- **Dismissed.** Wrong, missing context, or style preference\n\nPresent the explanation first (from Step 1), then the critique verdict below it. The explanation should stand on its own; someone who just wants to understand the system shouldn't wade through critique.\n","# How\n\nExplore the codebase to answer \"how does X work?\" questions. Produce clear architectural explanations at the level of a senior engineer onboarding onto a subsystem. Enough to build a working mental model, not annotated source code.\n\nTwo modes:\n\n1. **Explain** (default). Explore the codebase and produce a clear explanation\n2. **Critique.** Explain first, then run independent architectural reviews when the host permits them\n\n## Explain Mode\n\n### Step 1. Understand the Question and Assess Complexity\n\nParse what the user is asking about:\n\n- \"How does the rate limiter work?\", a subsystem\n- \"How do we handle billing for on-demand usage?\", a feature flow\n- \"How is the auth service structured?\", an architectural overview\n- \"Walk me through what happens when a user submits a form\", a runtime trace\n\nIdentify the scope. If ambiguous, state your best-guess interpretation before exploring. Don't ask. Let the user redirect if you're off.\n\n**Assess complexity to decide the approach:**\n\n- **Simple** (a single module, a small utility, a narrow question like \"how does function X work\"): skip explorer agents; the explainer explores and explains in a single pass. Go to Step 2b.\n- **Complex** (a subsystem spanning multiple files\u002Fservices, a cross-cutting feature, a full architectural overview): use parallel explorer agents when the user and host policy permit delegation; otherwise explore the same slices inline. Go to Step 2a.\n\nWhen in doubt, lean simple. You can widen the exploration if the first pass hits a wall.\n\n### Step 2a. Explore (complex questions only)\n\nDecompose the question into 2-4 parallel exploration angles, each a distinct slice of the subsystem so explorers don't duplicate work. Example split for \"how does the rate limiter work?\":\n\n- Explorer 1: data model and state management\n- Explorer 2: request path and enforcement\n- Explorer 3: configuration and metrics infrastructure\n\nThe right decomposition depends on the question. Use your judgment. Narrow questions: 2 explorers is fine. Broad subsystems: up to 4.\n\nWhen delegation is authorized, launch the explorers together with non-overlapping briefs, the inherited model, and read-only posture. Do not request a named model or broader permissions. When delegation is unavailable, work through the same exploration angles inline and keep their findings separate until synthesis.\n\nEach explorer gets the same base prompt from `references\u002Fexplorer-prompt.md` plus a specific exploration angle naming its slice. Each explorer should:\n- Start broad: Glob for relevant directories, Grep for key types\u002Finterfaces\u002Fclass names\n- Follow the thread: from an entry point, trace the call chain (callers, callees, data flow, type definitions)\n- Read the actual code, don't guess from file names\n- Stop when it can describe the full path from input to output (or trigger to effect) without hand-waving any step\n- Note things that are surprising, non-obvious, or that a newcomer would get wrong\n\nEach explorer returns structured findings: components found, flow traced, files read, anything non-obvious. Overlap between explorers is fine; the explainer reconciles.\n\nThen proceed to Step 3.\n\n### Step 2b. Direct Explain (simple questions)\n\nExplore and explain in one pass. Read `references\u002Fexplainer-prompt.md` for the communication style and output format. Use repository search and file reads directly; there is no need to delegate a narrow question.\n\nProceed to Step 4.\n\n### Step 3. Synthesize (complex questions only)\n\nSynthesize the exploration slices into one coherent explanation. You may use one independent explainer when delegation is already authorized and the subsystem is large; otherwise synthesize inline. Read `references\u002Fexplainer-prompt.md` for the full prompt template. Reconcile overlap and contradictions before presenting the result.\n\n### Step 4. Present\n\nPresent the explainer's output to the user. You may lightly edit for clarity or add context from the conversation, but don't substantially rewrite. The explainer's communication is the product.\n\n### Output Format\n\nFollow this structure, adapted to the question. Not every section is needed for every question.\n\n**Overview.** 1-2 paragraphs. What it is, what it does, why it exists. Enough to decide whether to keep reading.\n\n**Key Concepts.** The important types, services, or abstractions. Brief definition of each. Not exhaustive, just the ones needed to understand the rest.\n\n**How It Works.** The core of the explanation. Walk through the flow: what triggers it, what happens step by step, where data goes, the decision points. Prose, not pseudocode. Reference specific files and functions so the reader can go look, but don't dump code blocks unless a snippet is genuinely necessary.\n\n**Where Things Live.** A brief map of the relevant files\u002Fdirectories. Not every file, just the ones needed to start working in this area.\n\n**Gotchas.** Non-obvious or surprising things that would trip someone up. Historical context that explains why something looks weird. Known sharp edges.\n\n## Critique Mode\n\nTriggered when the user asks for architectural issues, problems, or improvements, not just understanding.\n\n### Step 1. Explain First\n\nRun the full explain flow above (Steps 1-4). You must understand the architecture before critiquing it.\n\n### Step 2. Run independent critiques\n\nAfter the explanation is complete, run independent architectural critiques. Use parallel critics only when the user and host policy permit delegation, and use the host's inherited model unless the user explicitly asks for model diversity. Otherwise perform one fresh critique pass after setting the explanation aside.\n\nRead `references\u002Fcritic-prompt.md` for the prompt template. Each critic gets:\n1. The explanation from Step 1 (so they don't re-explore)\n2. The relevant file paths (so they can read the actual code)\n3. The architectural critique rubric from `references\u002Fcritique-rubric.md`\n\n### Step 3. Lead Judgment\n\nSame framework as the interrogate skill. You're a pragmatic lead, not an aggregator.\n\nCategorize findings:\n- **Act on.** Architectural problems worth fixing now\n- **Consider.** Real concerns, but the cost\u002Fbenefit is unclear\n- **Noted.** Valid observations, low priority\n- **Dismissed.** Wrong, missing context, or style preference\n\nPresent the explanation first (from Step 1), then the critique verdict below it. The explanation should stand on its own; someone who just wants to understand the system shouldn't wade through critique.\n",{"name":5,"description":7},"main","6ef866230c1ba3459fd08fc65141dfd37fffd41a",0,"https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills","https:\u002F\u002Fskills.henryzh.dev","https:\u002F\u002Fgithub.com\u002Fcaezium","henryzh.dev","quant, data science, full-stack, robotics.\r\nbuilding cool stuff","2026-08-24 19:21:44",{"source":8,"skillName":5,"commands":54,"prompt":64,"zipAvailable":65},[55,58,61],{"tool":56,"command":57},"npx","npx skills add caezium\u002Fskills --skill how",{"tool":59,"command":60},"bunx","bunx skills add caezium\u002Fskills --skill how",{"tool":62,"command":63},"pnpm","pnpm dlx skills add caezium\u002Fskills --skill how","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002FSKILL.md",true,[67,73,77,81,85,89,93],{"path":68,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":71,"sourceHtmlUrl":72},"agents\u002Fopenai.yaml","REFERENCE","参考文档",93,"https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002Fagents\u002Fopenai.yaml",{"path":74,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":75,"sourceHtmlUrl":76},"evals\u002Fevals.json",671,"https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002Fevals\u002Fevals.json",{"path":78,"resourceType":78,"resourceTypeDesc":79,"sizeBytes":80,"sourceHtmlUrl":16},"LICENSE","许可证",1067,{"path":82,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":83,"sourceHtmlUrl":84},"references\u002Fcritic-prompt.md",2286,"https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002Freferences\u002Fcritic-prompt.md",{"path":86,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":87,"sourceHtmlUrl":88},"references\u002Fcritique-rubric.md",2877,"https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002Freferences\u002Fcritique-rubric.md",{"path":90,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":91,"sourceHtmlUrl":92},"references\u002Fexplainer-prompt.md",3121,"https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002Freferences\u002Fexplainer-prompt.md",{"path":94,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":95,"sourceHtmlUrl":96},"references\u002Fexplorer-prompt.md",2693,"https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fhow\u002Freferences\u002Fexplorer-prompt.md",[98,120,148,165,185,203],{"skillNo":99,"name":100,"nameZh":6,"description":101,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":102,"categories":103,"keywords":114,"publishedAt":39},"01M71JT4RYCAS18RF6003YYR1D","writing-for-agents","Writing documents for agents. Use when creating or editing skills, or modifying AGENTS.md or CLAUDE.md.","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fwriting-for-agents\u002FSKILL.md",[104,109],{"categoryNo":105,"categoryName":106,"taxonomyType":107,"taxonomyTypeDesc":108},"01JBSEEDCAT000000000000018","写作与文案","SCENARIO","使用场景",{"categoryNo":110,"categoryName":111,"taxonomyType":112,"taxonomyTypeDesc":113},"01JBSEEDCAT000000000000036","内容创作者","PROFESSION","适用职业",[115,36,116,117,33,118,119],"writing for agents","writing","editing","documents","when",{"skillNo":121,"name":122,"nameZh":6,"description":123,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":124,"categories":125,"keywords":129,"publishedAt":39},"01M71JT0GES4AZ5ADH8EADNN1Z","wrangler","Cloudflare Workers CLI for deploying, developing, and managing Workers, KV, R2, D1, Vectorize, Hyperdrive, Workers AI, Containers, Queues, Workflows, Pipelines, and Secrets Store. Load before running wrangler commands to ensure correct syntax and best practices. Biases towards retrieval from Cloudflare docs over pre-trained knowledge.","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fwrangler\u002FSKILL.md",[126],{"categoryNo":127,"categoryName":128,"taxonomyType":107,"taxonomyTypeDesc":108},"01JBSEEDCAT000000000000010","后端与接口",[122,130,131,132,133,134,135,136,137,138,139,140,141,36,142,143,144,145,146,147],"cloudflare","d1","knowledge","workflows","cli","secrets","commands","syntax","cloudflare workers","r2","queues","running","containers","correct","load","workers","best practices","store",{"skillNo":149,"name":150,"nameZh":6,"description":151,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":152,"categories":153,"keywords":154,"publishedAt":39},"01M71JSWX4R203M32K9A18AAW8","workers-best-practices","Reviews and authors Cloudflare Workers code against production best practices. Load when writing new Workers, reviewing Worker code, configuring wrangler.jsonc, or checking for common Workers anti-patterns (streaming, floating promises, global state, secrets, bindings, observability). Biases towards retrieval from Cloudflare docs over pre-trained knowledge.","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fworkers-best-practices\u002FSKILL.md",[],[155,145,146,156,157,158,36,130,159,132,160,122,116,144,119,32,161,162,163,164],"workers best practices","observability","reviews","promises","worker","new","common","streaming","state","anti-patterns",{"skillNo":166,"name":167,"nameZh":6,"description":168,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":169,"categories":170,"keywords":171,"publishedAt":39},"01M71JST6P88MXGT2MWDTPJGEG","wizard","Generate an interactive bash wizard that walks a human through steps only they can perform. Use when provisioning infrastructure, setting up credentials or CI secrets, walking an unfamiliar third-party dashboard, or running a one-off migration or cutover. Don't invoke this for steps the agent can perform itself.","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fwizard\u002FSKILL.md",[],[167,141,172,36,173,174,175,176,33,177,119,178,179,180,30,181,182,183,184,135],"setting","up","ci","provisioning infrastructure","only","invoke","Bash","one","interactive","credentials","dashboard","generate","human",{"skillNo":186,"name":27,"nameZh":6,"description":187,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":188,"sourceHtmlUrl":189,"categories":190,"keywords":191,"publishedAt":39},"01M71JSQ95S7BQ61DAPN95X2W4","Use for 'why does X work this way', 'why we picked Y', design rationale, regressions, postmortems, or data-backed thresholds. Discovers available evidence sources, searches each relevant category, and returns a cited read on decisions and tradeoffs. Use how for runtime behavior.","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fwhy\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fwhy\u002FSKILL.md",[],[27,192,193,194,195,36,35,196,33,197,20,5,198,199,200,201,202],"sources","thresholds","evidence","relevant","design","decisions","behavior","read","data","rationale","returns",{"skillNo":204,"name":205,"nameZh":6,"description":206,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":207,"categories":208,"keywords":212,"publishedAt":39},"01M71JSMHFVSH8HK8GXN6YAX6J","week-prep","Generate a briefing with calendar, tasks, and attention flags. Use for \"week prep\", \"briefing\", or \"what's on this week\".","https:\u002F\u002Fgithub.com\u002Fcaezium\u002Fskills\u002Fblob\u002Fmain\u002Fweek-prep\u002FSKILL.md",[209],{"categoryNo":210,"categoryName":211,"taxonomyType":107,"taxonomyTypeDesc":108},"01JBSEEDCAT000000000000023","Skill与智能体开发",[213,214,215,216,217,36,218,219,220,221,33,178,222,223,224,183,225,199,226,227],"week prep","prep","deep","attention","get","search","messages","briefing","MCP","tasks","with","slack","calendar","flags","grep"]