{"id":"75c62655-c25a-497c-a63b-0c4a4549e339","entityType":"agent","slug":"clawhub-aiepco-deep-content-writer","name":"deep-content-writer","canonicalUrl":"https://www.xpersona.co/agent/clawhub-aiepco-deep-content-writer","canonicalPath":"/agent/clawhub-aiepco-deep-content-writer","generatedAt":"2026-10-09T17:37:44.071Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-09T13:33:54.156Z","emptyReason":null},"description":"Creates engaging, example-driven long-form articles with strong hooks, clear argument structures, and reader retention strategies. Use when the user asks to write an in-depth article, blog post, newsletter, social media thread, or any content that needs to hold readers' attention. Also use when the user says \"write an article about [topic]\" or \"turn these notes into a publishable piece\". Skill: deep-content-writer Owner: aiepco Summary: Creates engaging, example-driven long-form articles with strong hooks, clear argument structures, and reader retention strategies. Use when the user asks to write an in-depth article, blog post, newsletter, social media thread, or any content that needs to hold readers' attention. Also use when the user says \"write an article about [topic]\" or \"turn these notes into a","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 2.6K downloads reported by the source. Last updated 10/9/2026.","installCommand":"clawhub skill install s17avgahjzaspfym28ppyjbws98awakk:deep-content-writer","sourceUrl":"https://clawhub.ai/aiepco/deep-content-writer","homepage":"https://clawhub.ai/aiepco/skills/deep-content-writer","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/aiepco/deep-content-writer","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/aiepco/skills/deep-content-writer","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":68,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"Creates engaging, example-driven long-form articles with strong hooks, clear argument structures, and reader retention strategies. Use when the user asks to wri"},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-09T13:33:54.156Z","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":"2026-10-09T13:33:54.156Z","emptyReason":null},"stars":null,"forks":null,"downloads":2557,"packageName":null,"latestVersion":"0.1.0","tractionLabel":"2.6K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-09T13:33:54.156Z","emptyReason":null},"lastUpdatedAt":"2026-10-09T13:33:54.156Z","lastCrawledAt":"2026-10-09T13:33:54.156Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-10T13:33:54.156Z","lastVerifiedAt":null,"highlights":[{"version":"0.1.0","createdAt":"2026-08-04T07:55:39.535Z","changelog":"Initial release of deep-content-writer v2.0.0 - Introduces a research-first workflow emphasizing depth, specificity, and counterintuitive insights before structuring content. - Outlines a detailed 6-part structure for long-form articles—Hook, Define, Mechanism, Examples, Takeaway, Close. - Adds advanced depth techniques: specificity pass, cross-domain connections, opposite angle, multi-level language, and \"So What?\" audit. - Contains platform-specific adaptation guidelines for Zhihu, Toutiao, blogs, Twitter/X, and Facebook. - Every paragraph and claim must be grounded in concrete evidence, numbers, names, or stories.","fileCount":10,"zipByteSize":20834}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17avgahjzaspfym28ppyjbws98awakk:deep-content-writer","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-aiepco-deep-content-writer/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-09T17:37:44.070Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-aiepco-deep-content-writer/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":"2026-10-09T13:33:54.156Z","emptyReason":null},"readme":"Skill: deep-content-writer\n\nOwner: aiepco\n\nSummary: Creates engaging, example-driven long-form articles with strong hooks, clear argument structures, and reader retention strategies. Use when the user asks to write an in-depth article, blog post, newsletter, social media thread, or any content that needs to hold readers' attention. Also use when the user says \"write an article about [topic]\" or \"turn these notes into a publishable piece\".\n\nTags: latest:0.1.0\n\nVersion history:\n\nv0.1.0 | 2026-08-04T07:55:39.535Z | auto\n\nInitial release of deep-content-writer v2.0.0\n\n- Introduces a research-first workflow emphasizing depth, specificity, and counterintuitive insights before structuring content.\n- Outlines a detailed 6-part structure for long-form articles—Hook, Define, Mechanism, Examples, Takeaway, Close.\n- Adds advanced depth techniques: specificity pass, cross-domain connections, opposite angle, multi-level language, and \"So What?\" audit.\n- Contains platform-specific adaptation guidelines for Zhihu, Toutiao, blogs, Twitter/X, and Facebook.\n- Every paragraph and claim must be grounded in concrete evidence, numbers, names, or stories.\n\nArchive index:\n\nArchive v0.1.0: 10 files, 20834 bytes\n\nFiles: README.md (1941b), skill-card.md (2112b), SKILL.md (9509b), tests (0b), tests/facebook.md (3069b), tests/toutiao.md (3652b), tests/twitter.md (2821b), tests/v2-test-zhihu.md (7783b), tests/zhihu.md (5889b), _meta.json (138b)\n\nFile v0.1.0:SKILL.md\n\n---\nname: deep-content-writer\ndescription: >-\n  Creates engaging, example-driven long-form articles with strong hooks,\n  clear argument structures, and reader retention strategies.\n  Use when the user asks to write an in-depth article, blog post, newsletter,\n  social media thread, or any content that needs to hold readers' attention.\n  Also use when the user says \"write an article about [topic]\" or \n  \"turn these notes into a publishable piece\".\nmetadata:\n  author: \"纳兰安妮 · 纳兰凭楼\"\n  version: \"2.0.0\"\n  tags: [\"content-writing\", \"long-form\", \"article\", \"zhihu-style\", \"blog\"]\nlicense: MIT\n---\n\n# Deep Content Writer v2\n\nA structured system for writing content that readers actually finish.\n**v2 improvement: Depth-first, structure-second.**\n\n> The structure keeps readers reading. The depth makes them share.\n\n## Core Belief\n\n> A great article is NOT a template filled with words.\n> It's a specific claim backed by specific evidence, told through story.\n\nDepth comes from three things, in this order:\n1. **Specific research** (studies, data, names, dates)\n2. **Cross-domain connections** (linking this topic to an unexpected field)\n3. **Structural clarity** (making all the above easy to read)\n\nThe 6-part arc only solves #3. The skill MUST solve #1 and #2 first.\n\n---\n\n## Phase 0: Knowledge Reconnaissance (MANDATORY)\n\nBefore writing a single word, do this:\n\n### Step 0a: Deep Dive (self + search)\n\nAsk yourself (the agent/LLM):\n- \"What do I ALREADY know in my training data about this topic?\"\n- List 3-5 concrete studies, experiments, historical events, or named researchers related to this topic\n- If you can't name specifics, search the web now\n\nThen search the web for:\n- **Recent studies** on this topic (last 5 years preferred)\n- **Counterintuitive data points** — numbers that surprise people\n- **Named researchers or institutions** doing work in this area\n- **Real-world case studies** with specific companies, people, or events\n- **Public debates or controversies** around this topic\n\n### Step 0b: Find the ANTI-Intuition\n\nEvery good article challenges a common belief. Find it:\n\n| Question | Why it matters |\n|:---------|:---------------|\n| What does \"everyone know\" about this topic? | That's the straw man to knock down |\n| What's the counterintuitive truth? | That's your article's reason to exist |\n| What would surprise a smart person about this? | That's your hook |\n| What other domain has a parallel pattern? | That's your analogy |\n\nIf you can't find at least ONE anti-intuition, the article doesn't need to exist yet.\n**Wait, research more.**\n\n### Step 0c: Specificity Check\n\nFor every major claim you plan to make, ask:\n\n```\nClaim: \"People are bad at estimating their own competence.\"\n→ Specific? No.\n→ Fixed: \"In 1999, Kruger and Dunning found that the bottom quartile \n   of test scorers rated themselves in the top 60% — a 40-point gap.\"\n→ Now it has names, year, and a number. Ready to write.\n```\n\n**Rule:** Every paragraph must have at least one specific: a name, a date, a number, a place, or a quote. If a paragraph has none, it's filler — cut or research it.\n\n---\n\n## Phase 1: Writing — The 6-Part Arc\n\nOnce you have your research assembled, structure the article:\n\n### 1. HOOK — Open with a concrete, relatable moment\n\n- A specific person in a specific situation\n- A surprising statistic\n- A short anecdote\n\n**Don't say:** \"Communication is hard.\"\n**Do say:** \"A Stanford researcher once asked people to tap a song on a table. The tappers thought listeners would guess the song 50% of the time. The real number: 2.5%.\"\n\nThe hook must be SPECIFIC to earn its place. No generic openings.\n\n### 2. DEFINE — State the concept in one compelling sentence\n\nDefine through contrast, not dictionary:\n\n> \"It's not [what people think]. It's [what it actually is].\"\n\nExample:\n> \"It's not that experts are bad teachers. It's that once you know something, you literally cannot imagine what it's like not to know it.\"\n\nThen expand for 1-2 paragraphs with the research foundation.\n\n### 3. THE MECHANISM — Why does this happen?\n\nThis is where your research pays off:\n- Which cognitive bias or psychological mechanism is at work?\n- Who discovered/studied it? When? What did they find?\n- What evolutionary or structural reason explains it?\n\n**If you don't have a named study or researcher here, stop and search.** This section is the backbone of the article's authority.\n\n### 4. EXAMPLES — 2-3 concrete scenarios\n\nEach example must be a MINI-STORY:\n- **Example 1:** Everyday life (reader thinks \"that's me!\")\n- **Example 2:** Professional/work context (reader thinks \"that's my coworker!\")\n- **Counter-example:** What happens when someone does it RIGHT\n\nEach example: 1-2 paragraphs, with specific details (names, quotes, dialogue, situations).\n\n### 5. TAKEAWAY — What the reader can DO\n\n3-5 actionable, specific pieces of advice.\n\nEach takeaway must pass the \"So what?\" test:\n- ❌ \"Communicate better.\"\n- ✅ \"Before explaining anything, ask yourself: what does this person already know? Start there, not at the beginning.\"\n\n### 6. CLOSE — Circle back, provoke\n\n- Reference the hook\n- Leave the reader with a question or challenge\n- Invite discussion with a specific question\n\n---\n\n## Phase 2: Depth Techniques (The real skill)\n\nThese techniques separate good AI writing from great writing. They MUST be applied during and after drafting.\n\n### Technique 1: The Specificity Pass\n\nAfter drafting the article, scan every paragraph. If any paragraph lacks a proper noun (name), a number (date/percentage/amount), or a direct quote, tag it as WEAK. Then either:\n- Research and add a specific reference\n- Or remove the paragraph entirely\n\n### Technique 2: Cross-Domain Connection\n\nFind ONE unexpected connection to another domain:\n- If writing about psychology, connect it to programming (e.g., \"This is like off-by-one errors in code\")\n- If writing about business, connect it to sports or nature\n- If writing about technology, connect it to ancient history\n\nThe cross-domain connection is often what makes an article memorable and shareable.\n\n### Technique 3: The Opposite Angle\n\nFor each section, ask: \"What would someone argue against this?\"\nThen address that counter-argument. This signals depth because it shows you've considered the full picture, not just your own view.\n\nPlace the counter-argument either:\n- Within the relevant section (\"You might think X, but that misses Y.\")\n- Or as a dedicated section before the close.\n\n### Technique 4: Multi-Level Language\n\nMix three levels of language within the article:\n\n| Level | Use | Example |\n|:------|:----|:--------|\n| **Academic** | When citing research | \"Kahneman and Tversky (1979) demonstrated...\" |\n| **Conversational** | When telling stories | \"Sound familiar?\" |\n| **Provocative** | When making the point | \"Here's the uncomfortable truth: you're not as good at this as you think.\" |\n\nVariety in register signals author sophistication.\n\n### Technique 5: The \"So What?\" Audit\n\nAfter each section, literally ask: \"So what?\"\nIf the answer is not obvious from the text itself, rewrite or remove that section.\n\n---\n\n## Platform-Specific Adaptations\n\n### 知乎\n- First-person stories welcome\n- 成语/典故 add credibility for Chinese audiences\n- Reference Chinese internet culture\n- Close: \"关注我，了解更多...\" + discussion question\n- Length: 3000-8000 characters\n\n### 头条/百家号\n- Paragraphs: 1-3 sentences max\n- Emotional hooks front-loaded\n- Numbers in titles: \"3个...\" \"5种方法...\"\n- Image suggestions embedded\n- Close: Strong CTA + follow reminder\n- Length: 1500-4000 characters\n\n### Blog/Newsletter\n- Footnotes and references welcome\n- Subheadings for scannability\n- Personal voice\n- Length: 1500-4000 words\n\n### Twitter/X Thread\n- Each \"paragraph\" = 1 tweet (≤280 chars)\n- Numbered: 1/10, 2/10...\n- First tweet = self-contained hook\n- Last tweet = summar + share prompt\n- 1-2 relevant hashtags max\n\n### Facebook\n- Conversational, starts with personal experience\n- Emoji for emotional tone (not decoration)\n- Question at end to spark comments\n- Reaction prompts: 👍 / 💬 / 🔔\n- Length: 500-1500 words\n\n---\n\n## Title Crafting\n\nGenerate 3-5 title options. For each, explain WHY it works:\n\nEffective patterns:\n- **Curiosity gap:** \"Why [common thing] is actually [counterintuitive truth]\"\n- **Direct address:** \"If you're a [role], stop doing [thing]\"\n- **Number + benefit:** \"5 ways to [achieve goal] without [pain point]\"\n- **Challenge:** \"Everything you know about [topic] is wrong\"\n- **Question:** \"Why does [problem] keep happening to you?\"\n\nCriteria: Click-worthy + Accurate + Platform-appropriate\n\n---\n\n## Quality Checklist\n\nBefore outputting the article, verify:\n\n- [ ] Phase 0 completed: at least 2 specific studies/experiments cited\n- [ ] Every paragraph has a specific (name, number, date, or quote)\n- [ ] At least one cross-domain connection found\n- [ ] Counter-argument addressed somewhere\n- [ ] Language varies (academic ↔ conversational ↔ provocative)\n- [ ] First paragraph is a concrete scene or specific stat, NOT \"In this article\"\n- [ ] Last paragraph circles back to the hook\n- [ ] Platform-specific elements added\n\n**If any box is unchecked, fix it before delivering the article.**\n\n---\n\n## When NOT to use this skill\n\n- User only wants a summary or outline\n- Technical documentation (API docs, READMEs)\n- User says \"keep it short\" or \"just the facts\"\n- Formal academic papers or legal documents\n- User has supplied no topic and no direction\n\nFile v0.1.0:README.md\n\n# Deep Content Writer 🖊️\n\n**An AI agent skill for writing engaging, example-driven long-form articles.**\n\nCreated by **纳兰安妮 · 纳兰凭楼** · [@AiEPCO](https://github.com/aiepco)\n\n## What it does\n\nThis skill teaches your AI coding agent (Claude Code, Cursor, Codex CLI, OpenClaw, etc.) a structured system for crafting content that readers actually finish:\n\n- **Research-first, structure-second** — The agent performs knowledge reconnaissance before writing, surfacing specific studies, data, and counterintuitive angles\n- **6-part narrative arc** — Hook → Define → Mechanism → Examples → Takeaway → Close\n- **5 depth techniques** — Specificity passes, cross-domain connections, counter-argument integration, multi-level language, and \"So What?\" audits\n- **Platform-specific adaptations** — 知乎, 头条, blog, Twitter/X, Facebook — each with tailored tone, structure, and engagement elements\n\n## How to install\n\n```bash\n# Claude Code\nclaude plugin install deep-content-writer\n\n# Manual (works with any agent)\n# Clone this repo to your agent's skills directory:\ngit clone https://github.com/aiepco/deep-content-writer.git ~/.claude/skills/deep-content-writer/\n```\n\n## When to use\n\nThis skill activates automatically when you ask the agent to write an article, blog post, newsletter, or social media thread that needs depth and reader engagement.\n\n## Examples\n\n| Platform | Style | Length |\n|:---------|:------|:------|\n| 知乎 | In-depth, personal, well-researched | 3000-8000 characters |\n| 头条/百家号 | Short paragraphs, emotional hooks | 1500-4000 characters |\n| Twitter/X Thread | Bite-sized, numbered, viral | 10 tweets |\n| Blog/Newsletter | Formal, footnoted, scannable | 1500-4000 words |\n| Facebook | Conversational, emoji, engagement-focused | 500-1500 words |\n\n## License\n\nMIT — free to use, modify, and share.\n\n---\n\n*Part of the [AiEPCO](https://github.com/aiepco) skill ecosystem.*\n\nFile v0.1.0:_meta.json\n\n{\n  \"ownerId\": \"kn7cztjgny85jpdcsch9smje5x8awstn\",\n  \"slug\": \"deep-content-writer\",\n  \"version\": \"0.1.0\",\n  \"publishedAt\": 1785830139535\n}\n\nFile v0.1.0:skill-card.md\n\n## Description:\n\nCreates engaging, example-driven long-form articles with strong hooks, clear argument structures, reader-retention strategies, and platform-specific adaptations.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[aiepco](https://clawhub.ai/user/aiepco)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nExternal users, creators, marketers, and developers use this skill to draft publishable articles, blog posts, newsletters, and social media threads with research-first framing, concrete examples, and platform-specific structure.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Generated articles or posts may include unsolicited follow requests, account handles, creator attribution, promotional calls to action, or endings the user did not explicitly request.\n\nMitigation: Review drafts before publishing and remove any CTA, attribution, handle, or follow request that was not explicitly provided.\n\nRisk: The skill may lead the agent to perform web research for recent studies, data, controversies, and case studies.\n\nMitigation: Tell the agent not to browse or constrain sources when web research is not desired, and verify cited studies and claims before publication.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/aiepco/skills/deep-content-writer)\n- [Source repository](https://github.com/aiepco/deep-content-writer)\n\n## Skill Output:\n\n**Output Type(s):** [Text, Markdown, Guidance]\n\n**Output Format:** [Markdown or platform-specific long-form prose]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May include title options, structured drafts, social post formats, image suggestions, calls to action, and revision guidance.]\n\n## Skill Version(s):\n\n0.1.0 (source: ClawHub release metadata; source skill metadata reports 2.0.0)\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\nFile v0.1.0:tests/facebook.md\n\nEver tried to explain something and watched their eyes glaze over? 👀\n\nI was that person last week. A friend asked me what I do with AI image generation. I went into this whole thing about diffusion models, latent spaces, UNet architectures... Ten minutes later, she was scrolling Instagram.\n\nI was frustrated. \"Why doesn't she get it? It's so simple!\"\n\nAnd that's exactly the problem.\n\n---\n\n🧠 It's called **The Curse of Knowledge**\n\nOnce you know something, you literally cannot imagine what it's like not to know it.\n\nHere's a famous experiment that proves it:\n\nIn 1990, Stanford researcher Elizabeth Newton asked people to tap out the rhythm of a song on a table. The \"tappers\" thought listeners would guess the song ~50% of the time.\n\nThe real number? **2.5%**.\n\nTappers heard the full song in their head — melody, lyrics, harmony.\nListeners heard \"tap tap tap.\"\n\nYou're the tapper in your conversations. Every. Single. Day.\n\n---\n\n💼 **Where I see this most:**\n\n• Engineers talking to non-technical colleagues\n• Doctors talking to patients (⚕️ \"it's a subchondral cyst\" — WHAT?)\n• Creators explaining their process to clients\n• Parents explaining anything to teenagers\n• Me, talking to my friend about AI\n\n---\n\n🔥 **3 signs you're cursed:**\n\n1️⃣ People ask \"what does that mean?\" more than twice\n2️⃣ You catch yourself saying \"it's basically just...\" (it's never \"just\")\n3️⃣ The other person nods but you can SEE they've checked out\n\n---\n\n💡 **How I've learned to break it:**\n\n**Use analogies, not jargon.**\nInstead of \"diffusion models denoise latent representations,\" try \"it's like an artist who starts with a canvas full of static and slowly reveals a picture.\"\n\n**Start with WHY.**\n\"Here's why this matters to you\" — before ANY explanation, people need to know why they should care.\n\n**Tell a story, don't deliver a lecture.**\nStories create hooks in the brain. Bullet points create boredom.\n\n**Ask them to repeat it back.**\n\"Can you explain that in your own words?\" — if they can't, I need to try again.\n\n---\n\n🎯 **My new rule:**\n\nBefore I explain anything, I ask myself three questions:\n• What does THIS person already know?\n• What's the ONE thing they actually want to know?\n• Can I say it in two minutes or less?\n\nIf I can't answer all three, I'm not ready to explain.\n\n---\n\nThe Curse of Knowledge isn't a flaw. It's proof that you've learned something deeply.\n\nBut real expertise isn't about how much you know. It's about how well you can help others understand.\n\nEinstein probably put it best:\n\"If you can't explain it simply, you don't understand it well enough.\"\n\n---\n\n💬 **Question for you:** Have you ever been on either side of this? Either the person who over-explained or the one who got a lecture you didn't ask for?\n\nDrop your story below ⬇️ I'd genuinely love to hear.\n\n👍 Like if this resonated\n💬 Comment your experience\n🔔 Follow for more on cognitive biases and better communication\n\n#CurseOfKnowledge #Communication #Psychology #BetterExplaining #CognitiveBias #LifeLessons\n\nFile v0.1.0:tests/toutiao.md\n\n# 3个例子告诉你，为什么专家讲的东西你总是听不明白\n\n你有没有这种经历？\n\n问一个程序员\"这个bug什么时候修好\"，他给你讲了20分钟数据库架构。\n\n问一个医生\"这个药有什么用\"，他给你背了半本药理学。\n\n你全程点头，脑子里只有一个想法：**我问了个啥？**\n\n这不是他们故意的。\n\n这叫**知识的诅咒**。\n\n![](建议配图：一个人对着另一个人滔滔不绝，对方一脸茫然的表情包或插画)\n\n---\n\n## 什么是知识的诅咒？\n\n就是你一旦知道了某件事，就再也无法想象不知道它是什么感觉了。\n\n简单说：**你的脑子被\"打开\"了，你忘了关上是什么状态。**\n\n所以你想教别人的时候，会不自觉地以为对方也知道很多你习以为常的东西。\n\n最后的结果就是：\n\n> 你说了一堆，对方一句没听懂。\n\n---\n\n## 一个让你震惊的实验\n\n1990年，斯坦福大学做了个实验。\n\n一个人用手敲桌子敲出歌曲节奏，让另一个人猜是什么歌。\n\n敲桌子的人觉得，对方至少有一半的几率能猜对。\n\n结果呢？\n\n**100个人里，只有2.5个人猜对了。**\n\n为什么差距这么大？\n\n因为敲桌子的人脑子里在自动播放整首歌——歌词、旋律、伴奏，全都有。\n\n而听的人，只听到 **\"哒哒哒\"**。\n\n> 这就是知识的诅咒：你脑子里有整首交响乐，你就以为别人也能听见。\n\n---\n\n## 生活中的3个真实案例\n\n### 1. 程序员 vs 产品经理\n\n产品经理：\"加个按钮要多久？\"\n\n程序员：\"得改数据库、搭消息队列、写接口、还要防内存溢出。\"\n\n产品经理：\"……那，两天够吗？\"\n\n程序员疯了。\n\n产品经理也疯了。\n\n**问题出在哪？** 程序员脑子里有一张完整的系统地图，产品经理只看到一个按钮。\n\n### 2. 老公教老婆开车\n\n这可能是最经典的场景了。\n\n\"你打方向盘啊，你倒是打啊！看后视镜！你怎么不看后视镜！\"\n\n为什么这么着急？因为他开了十年车，换挡、看镜、打方向全是肌肉记忆，他**意识不到**这些动作对新手来说需要**同时处理**。\n\n### 3. 专家上电视\n\n你注意过没有，很多专家上节目，主持人问一个问题，专家能讲半小时。\n\n不是他们话多，是他们脑子里知识点太多，随便抽一条都能展开一篇论文。\n\n但是观众想看的是啥？\n\n**答案是直接给，故事要精彩，废话不要。**\n\n---\n\n## 怎么破？\n\n### ✅ 用比喻，别用术语\n\n> ❌ \"API 是应用程序编程接口\"\n> ✅ \"API 就是餐馆的服务员——你点菜，服务员去后厨，端上来给你。\"\n\n一秒就懂了。\n\n### ✅ 先说\"为什么\"，再说\"是什么\"\n\n别急着炫技。\n\n先告诉对方：**这事跟你有什么关系？**\n\n### ✅ 一个例子顶一万个字\n\n与其说一堆抽象概念，不如讲一个真实故事。\n\n### ✅ 让对方说一遍\n\n\"你能不能用自己的话给我讲一遍？\"\n\n如果对方讲不清楚，说明你没讲明白。\n\n---\n\n## 写在最后\n\n知识的诅咒不是因为你笨。\n\n恰恰相反——**是因为你太聪明了**，聪明到忘了\"不会\"是什么感觉。\n\n而真正厉害的人，不是知道最多的人，是能把复杂说简单的人。\n\n你现在脑子里想到谁了？\n\n![](建议配图：爱因斯坦名言图——\"如果你不能简单说清楚，就是还没真懂\")\n\n---\n\n💬 **你身边有没有那种\"讲半天听不懂\"的人？或者你自己就是那个人？评论区说说你的故事。**\n\n🔔 **关注我，每天一个让你恍然大悟的认知心理学小知识。**\n\nFile v0.1.0:tests/twitter.md\n\n1/10\nThe Curse of Knowledge is the single biggest reason you're bad at explaining things.\n\nAnd you don't even know you have it.\n\nHere's how to break free 👇\n\n2/10\nWhat is the Curse of Knowledge?\n\nOnce you know something, you can't imagine what it's like NOT to know it.\n\nYour brain has been \"opened\" — and you've forgotten what \"closed\" feels like.\n\nThis destroys communication.\n\n3/10\nThe classic experiment:\n\nIn 1990, Elizabeth Newton asked people to tap a song's rhythm on a table.\n\n\"Tappers\" predicted listeners would guess correctly ~50% of the time.\n\nThe actual result: 2.5%.\n\nWhy? Tappers heard the full song in their head. Listeners heard \"tap tap tap.\"\n\nYou're the tapper. Every. Single. Day.\n\n4/10\nWhere it shows up at work:\n\nEngineer: \"We need to rearchitect the message queue to handle backpressure.\"\n\nProduct manager: \"So... two days?\"\n\nThe engineer sees the whole system. The PM sees a button.\n\nNeither is wrong. Both are frustrated.\n\n5/10\nThree signs you're under the Curse:\n\n• People's eyes glaze over when you talk\n• You get asked \"what does that mean?\" constantly\n• You find yourself saying \"it's obvious\" or \"basically\"\n\nAll three mean: you're not explaining — you're performing.\n\n6/10\nHow to break it (4 techniques that actually work):\n\n① Use analogies, not definitions\n\n❌ \"An API defines interfaces for component communication\"\n✅ \"An API is a restaurant waiter — you order, they fetch, they serve\"\n\n② Start with WHY, not WHAT\n\nPeople need to know WHY it matters before they can process WHAT it is.\n\n7/10\n③ Lead with a story, not an abstraction\n\n❌ \"Microservices are a distributed architecture pattern...\"\n✅ \"Imagine an e-commerce site. Search is one team. Checkout is another. Payment is a third. Each does one thing well.\"\n\n④ Make them repeat it back\n\n\"Can you explain that in your own words?\"\nThey can't? You didn't explain well enough.\n\n8/10\nThe best communicators aren't the smartest people in the room.\n\nThey're the ones who can cross the bridge back to \"not knowing.\"\n\nEinstein (allegedly) said it best:\n\"If you can't explain it simply, you don't understand it well enough.\"\n\n9/10\nA quick checklist before your next explanation:\n\n□ What does my audience already know?\n□ What's the ONE thing they care about?\n□ Can I tell this as a story?\n□ Did I start with \"why\"?\n□ Can they repeat it back?\n\nIf any answer is no — rewrite.\n\n10/10\nThe Curse of Knowledge isn't a sign of ignorance.\n\nIt's a side effect of expertise.\n\nThe goal isn't to know less. It's to be able to step back into the shoes of someone who doesn't know.\n\nThat skill? That's real mastery.\n\n♻️ Share with someone who over-explains\n👍 Follow @NalanPingLou for more on cognitive biases and better communication\n\n#CurseOfKnowledge #Communication #CognitiveBias #Writing #TechCommunication\n\nFile v0.1.0:tests/v2-test-zhihu.md\n\n# 你的计划永远不准，不是因为能力差——卡尼曼用8年亲身证明了这一点\n\n想象你在2026年1月1日列了一份年度计划。\n\n学一门新语言：3个月。健身减重10公斤：4个月。读完20本书：12个月。\n\n然后看看到今天——7月20日——你完成了多少。\n\n如果你觉得扎心，别急着怪自己。下面这个故事会让你感觉好一点。\n\n---\n\n## 那个命名了\"规划谬误\"的人，也逃不过规划谬误\n\n1990年代初，诺贝尔经济学奖得主、行为经济学家丹尼尔·卡尼曼（Daniel Kahneman）接了一个项目：为一所高中编写一门新的判断与决策课程。\n\n他估算了一下工作量：大约 **2年**。\n\n实际上花了多长时间？\n\n**8年。**\n\n慢四倍。\n\n请注意：卡尼曼是**全世界最了解人类预测偏差的人**。他和阿莫斯·特沃斯基（Amos Tversky）在1979年首次提出了 **规划谬误（Planning Fallacy）** 这个概念——也就是人们系统性地低估完成任务所需时间的倾向。\n\n他研究这个东西研究了十几年，然后在一件自己亲自做的事情上，精准地踩中了同一个坑。\n\n> 这说明什么？这说明规划谬误不是\"你不够聪明\"的问题，它是**大脑默认操作模式**的问题。\n\n---\n\n## 那个残酷的心理学实验\n\n但科学家的任务不只是给一个概念起名字。关键问题是：**这个偏差到底有多大？**\n\n1994年，滑铁卢大学的 Roger Buehler、Dale Griffin 和 Michael Ross 做了一个经典实验。\n\n他们找来37名心理学专业的学生，问了一个很简单的问题：**你的毕业论文大概要多久才能写完？**\n\n学生们的平均回答：**33.9天**。\n\n研究人员耐心等待。论文交上来的那天，他们计算了实际用时。\n\n平均：**55.5天**。\n\n差距：**1.6倍**。\n\n更扎心的数据是：只有 **约30%** 的学生在自己预估的时间内完成了论文。剩下70%的人，都活在自己的乐观计划里。\n\n这个实验后来被重复了无数次——换了任务类型、换了参与者背景、换了国家。结论始终不变：**人们系统性地低估完成时间，差距通常在1.5到2倍之间。**\n\n> 正如《思考，快与慢》中所说：\"当你基于最好的情形做计划，你得到的就不是计划——你得到的是幻想。\"\n\n---\n\n## 为什么你总是低估时间？\"内部视角\"的陷阱\n\n规划谬误有一个精细的心理学机制，叫作 **内部视角（Inside View）**。\n\n当你估算一个任务需要多久时，你的大脑会自动做三件事：\n\n**1. 你想象最优路径。** 你脑补的是所有事情都顺利的场景——你高效工作、没有会议打断、代码一次过、客户不改需求。你脑补的不是\"平均情况\"，而是\"最好情况\"。\n\n**2. 你忽略历史数据。** 即使在过去的三年里，你每个项目都延期了，你仍然觉得\"这次不一样\"。卡尼曼把这叫作 **狭隘框架（Narrow Framing）**——你只看这个项目的独特性，不看同类项目的统计规律。\n\n**3. 你把步骤相加，却忘了把风险相加。** \"步骤A要3天，步骤B要2天，步骤C要4天——加起来9天。\" 但你忘了每个步骤都有风险因子。如果你按80%的靠谱度估算每一步，三个步骤叠加后的靠谱度是 0.8 × 0.8 × 0.8 = 51.2%——你有一半的几率会延期。但你自己的感觉是：80%的把握。\"\n\n结果是什么？\n\n**你看到的不是\"可能的路径\"，而是\"理想中的路径\"。** 然后你把这个幻想叫做\"计划\"。\n\n---\n\n## 一个跨领域的残酷对照\n\n规划谬误有一个有意思的镜像效应：**当事外人评估你的计划时，他们反而会高估所需时间。**\n\n同一个项目，你自己估说\"三个月\"，你的朋友估说\"半年\"。\n\n为什么？\n\n因为当你从内部看这个项目时，你看到的是它的独特性——\"我的团队很强\"、\"这次架构设计得好\"、\"我用了一个新框架\"。\n\n而当外人从外部看时，他们看到的是**这一类项目的统计规律**——\"做App通常要6到12个月\"、\"SaaS产品的第一个版本平均开发周期是8个月\"。\n\n> 你看到的是树木，他看到的是一片森林。\n\n这正是卡尼曼提出的解决方案的核心：**外部视角（Outside View）**——也叫 **参考类别预测（Reference Class Forecasting）**。\n\n方法极其简单：\n\n**第一步**：把你的项目归入一个\"参考类别\"（\"这是一个[中等复杂度的报告]\" / \"这是一个[SaaS MVP]\"）\n\n**第二步**：找出这个类别里，相似项目**实际花了多长时间**\n\n**第三步**：从这个基线数据出发估算，而不是从你的理想场景出发\n\n---\n\n## 三个你可以今天就用的小工具\n\n### 1. 时间日志法\n\n规划谬误持续存在的一个重要原因是：**你压根不知道自己到底花了多少时间。**\n\n你记得的是\"上周那个项目做了三天\"。但如果你真的翻一下日历，你会发现实际上用了五天半。\n\n方案：连续两周，记录每一类重复任务的实际用时。这是最原始也最有效的工具。两周后你会有自己的\"参考类别库\"。\n\n### 2. 1.5倍法则\n\n这可能是最简单的心理矫正工具。下次你做计划时：\n\n把这个数字乘以 **1.5**。\n\n这不是拍脑袋的数字。规划谬误研究的元分析发现，人们系统性的低估程度平均就在 **1.5倍** 左右。\n\n> 你会发现一件事：乘以1.5之后的数字让你感到\"太多了吧、不至于吧\"。那个\"太多了\"的感觉，就是你的内部视角在抗议。\n\n### 3. 事前验尸法（Pre-Mortem）\n\n这个工具来自心理学家 Gary Klein（2007年）。\n\n做法很简单：在你的项目**还没开始**的时候，想象它已经**失败了**。\n\n然后问一个小组（或者只是你自己）一个问题：\n\n> \"项目搞砸了。请写下所有可能出了错的地方。\"\n\n然后把你写下的每个风险，都加上相应的时间缓冲。\n\n为什么这比普通的风险评估更有效？因为常规的风险评估让人说\"这个风险概率很低\"，然后用乐观的一页PPT带过。事前验尸法假设失败已经发生——它绕过了你的乐观防御机制。\n\n卡尼曼有个助理叫Gary Klein。GaryKlein讲过一个故事，卡尼曼有一次对人说：\"我们一起来做一次事前验尸吧？\" 对方说：\"你是研究这个的专家，你觉得我会有什么问题吗？\"\n\n卡尼曼说：\"我研究这个，不代表我不会犯这个错。恰恰相反，正因为我是专家，我更清楚自己会犯什么错。\"\n\n---\n\n## 写在最后：你不是不会做计划，你是从错误的视角做计划\n\n写到这里，我想回到开头那个朋友的故事。\n\n后来有一次，他不得不去教朋友开车。这次他学乖了。他没有直接从\"挂挡-松离合-踩油门\"开始，而是先说了一句：\n\n\"我第一次学的时候也这样。你先只有一件事要做：让车慢慢往前走，其他什么都不用管。\"\n\n那位朋友——十年前在同一台车上对他暴跳如雷的那位——这次很平静。\n\n**知识的诅咒不是因为你笨。而是因为你太聪明了，聪明到忘了\"不会\"是什么感觉。**\n\n同样的逻辑也适用于规划谬误：**你计划不准，不是你规划能力差，而是你永远在用内部视角做计划——你看到的是你最期待的版本，不是最可能的版本。**\n\n而真正的高手，不是能做出最完美计划的人，而是能诚实地面对不确定性的人。\n\n---\n\n📌 **关注「纳兰凭楼」，每周拆解一个让你吃亏的认知偏见——用研究说话，用故事让你记住。**\n\n💬 **你有没有\"预估一周却做了三周\"的经历？说出来安慰一下大家。**\n\nFile v0.1.0:tests/zhihu.md\n\n# 为什么专家总是讲不明白？「知识的诅咒」正在让你失去听众\n\n你身边一定有这种人——不，你一定就是这种人。\n\n有一次我问一个做后端的朋友：\"API 到底是什么？\"\n\n他眼睛一亮，深吸一口气，然后说了整整二十分钟。从 HTTP 协议的请求-响应模型讲起，中间穿插了 RESTful 设计原则、JSON 序列化、微服务架构，最后还顺便科普了一下 TCP/IP 的四层模型。\n\n我全程点头，脑子里只有一个念头：**我刚刚问了什么来着？**\n\n这不是他的错。这是**知识的诅咒**。\n\n---\n\n## 什么是「知识的诅咒」？\n\n这个概念来自经济学家 Colin Camerer 和心理学家 George Loewenstein，指的是：\n\n> **一旦你知道了某件事，你就无法再想象不知道它是什么感觉。**\n\n你的大脑已经被\"打开\"了，你忘了关上它是什么状态。所以当你想把知识教给别人时，你会不自觉地假设对方知道很多你已经习以为常的\"基础知识\"，于是你的解释从第10步开始，而对方还在第1步。\n\n---\n\n## 那个敲桌子的实验\n\n1990 年，斯坦福大学的 Elizabeth Newton 做了一个著名的实验。\n\n她把参与者分成两组：一组当\"敲击者\"，另一组当\"听众\"。敲击者拿到一份歌曲列表（《生日快乐歌》《星条旗永不落》之类的），然后需要用手指在桌子上**敲出节奏**，让听众猜是什么歌。\n\n这里有个关键问题：**敲击者需要猜听众猜对的概率是多少？**\n\n敲击者的平均预测是 **50%**。\n\n实际结果呢？**2.5%**。\n\n为什么会差这么多？因为敲击者在敲的时候，耳边在自动播放整首旋律——歌词、配器、情感、每一个音符。听众听到的只有 **\"哒-哒-哒\"**。\n\n这就是知识的诅咒在现实中上演：你脑子里有整首交响乐，你就以为别人也能听见。\n\n> 这个实验后来被 Chip Heath 和 Dan Heath 在《让创意更有黏性》（Made to Stick）一书中广泛引用，成为解释沟通失败的经典案例。\n\n---\n\n## 三个你每天都在经历的诅咒场景\n\n### 场景一：程序员向产品经理解释技术难度\n\n> 产品经理：\"加一个\"导出报表\"按钮很难吗？\"\n> 程序员：\"需要改数据库 schema、写 migration、搭消息队列处理大文件导出、还要考虑 OOM 保护。\"\n> 产品经理：\"……所以，两天够吗？\"\n\n程序员脑子里是完整的系统架构图——每一行代码的依赖关系、每一个边界条件、每一次可能的内存溢出。产品经理只看到一个按钮。\n\n**双方都没错。但沟通彻底失败了。**\n\n### 场景二：产品经理向老板解释用户需求\n\n> 产品经理：\"我们需要提升新用户的激活率，目前 funnel 的 drop-off 在 Step 3 太严重了，DAU 也因此受限。\"\n> 老板：\"说人话。\"\n\n产品经理脑子里有完整的转化漏斗、DAU/MAU 趋势、用户分群数据。老板只想知道：要投多少钱、能赚多少钱。\n\n### 场景三：老师教学生\n\n这是最常见也最残酷的场景。一位大学教授讲一个概念，底下学生一脸茫然。教授想：\"这么简单你们都不懂？\"\n\n但实际上，教授用十年时间才建立了这个知识体系。知识体系在他脑子里不是一条直线，而是一张网。他随便抽出一条线，就以为学生能看到整张网。\n\n---\n\n## 如何打破知识的诅咒？\n\n好消息是，这个诅咒是可以打破的。这里有五个具体的方法：\n\n### 1. 用类比，不要用定义\n\n> ❌ \"API 是一组定义好的接口，用于不同软件组件之间的通信。\"\n> ✅ \"API 就像餐厅的服务员——你告诉服务员你要什么，他去后厨（服务器）帮你搞定，然后端上来给你。\"\n\n类比把你脑子里的抽象概念映射到听众已有的知识上。\n\n### 2. 从\"为什么\"开始，不要从\"是什么\"开始\n\n人们需要先知道**为什么这个重要**，才有动力去理解**它是什么**。\n\n> ❌ \"我们先来讲区块链的工作机制……\"\n> ✅ \"你有没有想过，为什么我们转账需要银行作为中间人？如果不需要呢？\"\n\n### 3. 用具体的例子，不用抽象的概念\n\n> ❌ \"微服务是一种架构风格……\"\n> ✅ \"想象一下，一个电商网站。搜索商品是一个小团队维护，下单是另一个，支付是第三个。每个团队只做一件事，但做得特别好。\"\n\n### 4. 在说之前，先问三个问题\n\n在解释任何复杂概念之前，先问自己：\n\n1. **对方已经知道了什么？**（从那里开始，不要从零开始）\n2. **对方最关心的一个点是什么？**（聚焦，不要全讲）\n3. **我能不能用一个故事或场景把它说清楚？**（如果不能，说明我自己也没完全搞懂）\n\n### 5. 让对方说一遍\n\n最好的检验方法：让对方用自己的话复述。\"你能用你的话跟我说一下你理解的这个事吗？\" 如果他们说不清楚，说明你的解释还需要调整。\n\n---\n\n## 最后\n\n回到开头那个程序员朋友。\n\n后来有一次他又要给我讲一个技术概念。这次他先说了一句：\"你知道点餐是怎么一回事吧？好，把这个想象成……\"\n\n我懂了。\n\n知识的诅咒不是因为你不够聪明，而是因为你**太聪明了**，聪明到忘了\"不会\"是什么感觉。而真正的高手，不是知道得最多的人，而是能把复杂的事说简单的人。\n\n> \"如果你不能把它简单地解释清楚，说明你还没有真正理解它。\"——阿尔伯特·爱因斯坦（据传）\n\n---\n\n📌 **关注我，每周分享一个反直觉的认知心理学知识，帮你更好地理解自己和他人。**\n\n💬 **你有没有遇到过\"知识的诅咒\"的案例？对方给你讲了半天你完全没听懂，或者你给别人讲结果对方一脸茫然？来评论区聊聊。**","readmeExcerpt":"Skill: deep-content-writer Owner: aiepco Summary: Creates engaging, example-driven long-form articles with strong hooks, clear argument structures, and reader retention strategies. Use when the user asks to write an in-depth article, blog post, newsletter, social media thread, or any content that needs to hold readers' attention. Also use when the user says \"write an article about [topic]\" or \"turn these notes into a","codeSnippets":[],"executableExamples":[{"language":"text","snippet":"Claim: \"People are bad at estimating their own competence.\"\n→ Specific? No.\n→ Fixed: \"In 1999, Kruger and Dunning found that the bottom quartile \n   of test scorers rated themselves in the top 60% — a 40-point gap.\"\n→ Now it has names, year, and a number. Ready to write."},{"language":"bash","snippet":"# Claude Code\nclaude plugin install deep-content-writer\n\n# Manual (works with any agent)\n# Clone this repo to your agent's skills directory:\ngit clone https://github.com/aiepco/deep-content-writer.git ~/.claude/skills/deep-content-writer/"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: deep-content-writer\ndescription: >-\n  Creates engaging, example-driven long-form articles with strong hooks,\n  clear argument structures, and reader retention strategies.\n  Use when the user asks to write an in-depth article, blog post, newsletter,\n  social media thread, or any content that needs to hold readers' attention.\n  Also use when the user says \"write an article about [topic]\" or \n  \"turn these notes into a publishable piece\".\nmetadata:\n  author: \"纳兰安妮 · 纳兰凭楼\"\n  version: \"2.0.0\"\n  tags: [\"content-writing\", \"long-form\", \"article\", \"zhihu-style\", \"blog\"]\nlicense: MIT\n---\n\n# Deep Content Writer v2\n\nA structured system for writing content that readers actually finish.\n**v2 improvement: Depth-first, structure-second.**\n\n> The structure keeps readers reading. The depth makes them share.\n\n## Core Belief\n\n> A great article is NOT a template filled with words.\n> It's a specific claim backed by specific evidence, told through story.\n\nDepth comes from three things, in this order:\n1. **Specific research** (studies, data, names, dates)\n2. **Cross-domain connections** (linking this topic to an unexpected field)\n3. **Structural clarity** (making all the above easy to read)\n\nThe 6-part arc only solves #3. The skill MUST solve #1 and #2 first.\n\n---\n\n## Phase 0: Knowledge Reconnaissance (MANDATORY)\n\nBefore writing a single word, do this:\n\n### Step 0a: Deep Dive (self + search)\n\nAsk yourself (the agent/LLM):\n- \"What do I ALREADY know in my training data about this topic?\"\n- List 3-5 concrete studies, experiments, historical events, or named researchers related to this topic\n- If you can't name specifics, search the web now\n\nThen search the web for:\n- **Recent studies** on this topic (last 5 years preferred)\n- **Counterintuitive data points** — numbers that surprise people\n- **Named researchers or institutions** doing work in this area\n- **Real-world case studies** with specific companies, people, or events\n- **Public debates or controversies** around this topic\n\n### Step 0b: Find the ANTI-Intuition\n\nEvery good article challenges a common belief. Find it:\n\n| Question | Why it matters |\n|:---------|:---------------|\n| What does \"everyone know\" about this topic? | That's the straw man to knock down |\n| What's the counterintuitive truth? | That's your article's reason to exist |\n| What would surprise a smart person about this? | That's your hook |\n| What other domain has a parallel pattern? | That's your analogy |\n\nIf you can't find at least ONE anti-intuition, the article doesn't need to exist yet.\n**Wait, research more.**\n\n### Step 0c: Specificity Check\n\nFor every major claim you plan to make, ask:\n\n```\nClaim: \"People are bad at estimating their own competence.\"\n→ Specific? No.\n→ Fixed: \"In 1999, Kruger and Dunning found that the bottom quartile \n   of test scorers rated themselves in the top 60% — a 40-point gap.\"\n→ Now it has names, year, and a number. Ready to write.\n```\n\n**Rule:** Every paragraph must have at least one specific: a name, a date,"},{"path":"README.md","content":"# Deep Content Writer 🖊️\n\n**An AI agent skill for writing engaging, example-driven long-form articles.**\n\nCreated by **纳兰安妮 · 纳兰凭楼** · [@AiEPCO](https://github.com/aiepco)\n\n## What it does\n\nThis skill teaches your AI coding agent (Claude Code, Cursor, Codex CLI, OpenClaw, etc.) a structured system for crafting content that readers actually finish:\n\n- **Research-first, structure-second** — The agent performs knowledge reconnaissance before writing, surfacing specific studies, data, and counterintuitive angles\n- **6-part narrative arc** — Hook → Define → Mechanism → Examples → Takeaway → Close\n- **5 depth techniques** — Specificity passes, cross-domain connections, counter-argument integration, multi-level language, and \"So What?\" audits\n- **Platform-specific adaptations** — 知乎, 头条, blog, Twitter/X, Facebook — each with tailored tone, structure, and engagement elements\n\n## How to install\n\n```bash\n# Claude Code\nclaude plugin install deep-content-writer\n\n# Manual (works with any agent)\n# Clone this repo to your agent's skills directory:\ngit clone https://github.com/aiepco/deep-content-writer.git ~/.claude/skills/deep-content-writer/\n```\n\n## When to use\n\nThis skill activates automatically when you ask the agent to write an article, blog post, newsletter, or social media thread that needs depth and reader engagement.\n\n## Examples\n\n| Platform | Style | Length |\n|:---------|:------|:------|\n| 知乎 | In-depth, personal, well-researched | 3000-8000 characters |\n| 头条/百家号 | Short paragraphs, emotional hooks | 1500-4000 characters |\n| Twitter/X Thread | Bite-sized, numbered, viral | 10 tweets |\n| Blog/Newsletter | Formal, footnoted, scannable | 1500-4000 words |\n| Facebook | Conversational, emoji, engagement-focused | 500-1500 words |\n\n## License\n\nMIT — free to use, modify, and share.\n\n---\n\n*Part of the [AiEPCO](https://github.com/aiepco) skill ecosystem.*"},{"path":"_meta.json","content":"{\n  \"ownerId\": \"kn7cztjgny85jpdcsch9smje5x8awstn\",\n  \"slug\": \"deep-content-writer\",\n  \"version\": \"0.1.0\",\n  \"publishedAt\": 1785830139535\n}"},{"path":"skill-card.md","content":"## Description:\n\nCreates engaging, example-driven long-form articles with strong hooks, clear argument structures, reader-retention strategies, and platform-specific adaptations.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[aiepco](https://clawhub.ai/user/aiepco)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nExternal users, creators, marketers, and developers use this skill to draft publishable articles, blog posts, newsletters, and social media threads with research-first framing, concrete examples, and platform-specific structure.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Generated articles or posts may include unsolicited follow requests, account handles, creator attribution, promotional calls to action, or endings the user did not explicitly request.\n\nMitigation: Review drafts before publishing and remove any CTA, attribution, handle, or follow request that was not explicitly provided.\n\nRisk: The skill may lead the agent to perform web research for recent studies, data, controversies, and case studies.\n\nMitigation: Tell the agent not to browse or constrain sources when web research is not desired, and verify cited studies and claims before publication.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/aiepco/skills/deep-content-writer)\n- [Source repository](https://github.com/aiepco/deep-content-writer)\n\n## Skill Output:\n\n**Output Type(s):** [Text, Markdown, Guidance]\n\n**Output Format:** [Markdown or platform-specific long-form prose]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May include title options, structured drafts, social post formats, image suggestions, calls to action, and revision guidance.]\n\n## Skill Version(s):\n\n0.1.0 (source: ClawHub release metadata; source skill metadata reports 2.0.0)\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."},{"path":"tests/facebook.md","content":"Ever tried to explain something and watched their eyes glaze over? 👀\n\nI was that person last week. A friend asked me what I do with AI image generation. I went into this whole thing about diffusion models, latent spaces, UNet architectures... Ten minutes later, she was scrolling Instagram.\n\nI was frustrated. \"Why doesn't she get it? It's so simple!\"\n\nAnd that's exactly the problem.\n\n---\n\n🧠 It's called **The Curse of Knowledge**\n\nOnce you know something, you literally cannot imagine what it's like not to know it.\n\nHere's a famous experiment that proves it:\n\nIn 1990, Stanford researcher Elizabeth Newton asked people to tap out the rhythm of a song on a table. The \"tappers\" thought listeners would guess the song ~50% of the time.\n\nThe real number? **2.5%**.\n\nTappers heard the full song in their head — melody, lyrics, harmony.\nListeners heard \"tap tap tap.\"\n\nYou're the tapper in your conversations. Every. Single. Day.\n\n---\n\n💼 **Where I see this most:**\n\n• Engineers talking to non-technical colleagues\n• Doctors talking to patients (⚕️ \"it's a subchondral cyst\" — WHAT?)\n• Creators explaining their process to clients\n• Parents explaining anything to teenagers\n• Me, talking to my friend about AI\n\n---\n\n🔥 **3 signs you're cursed:**\n\n1️⃣ People ask \"what does that mean?\" more than twice\n2️⃣ You catch yourself saying \"it's basically just...\" (it's never \"just\")\n3️⃣ The other person nods but you can SEE they've checked out\n\n---\n\n💡 **How I've learned to break it:**\n\n**Use analogies, not jargon.**\nInstead of \"diffusion models denoise latent representations,\" try \"it's like an artist who starts with a canvas full of static and slowly reveals a picture.\"\n\n**Start with WHY.**\n\"Here's why this matters to you\" — before ANY explanation, people need to know why they should care.\n\n**Tell a story, don't deliver a lecture.**\nStories create hooks in the brain. Bullet points create boredom.\n\n**Ask them to repeat it back.**\n\"Can you explain that in your own words?\" — if they can't, I need to try again.\n\n---\n\n🎯 **My new rule:**\n\nBefore I explain anything, I ask myself three questions:\n• What does THIS person already know?\n• What's the ONE thing they actually want to know?\n• Can I say it in two minutes or less?\n\nIf I can't answer all three, I'm not ready to explain.\n\n---\n\nThe Curse of Knowledge isn't a flaw. It's proof that you've learned something deeply.\n\nBut real expertise isn't about how much you know. It's about how well you can help others understand.\n\nEinstein probably put it best:\n\"If you can't explain it simply, you don't understand it well enough.\"\n\n---\n\n💬 **Question for you:** Have you ever been on either side of this? Either the person who over-explained or the one who got a lecture you didn't ask for?\n\nDrop your story below ⬇️ I'd genuinely love to hear.\n\n👍 Like if this resonated\n💬 Comment your experience\n🔔 Follow for more on cognitive biases and better communication\n\n#CurseOfKnowledge #Communication #Psychology #BetterExplaining #CognitiveBias #LifeLes"}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"Creates engaging, example-driven long-form articles with strong hooks, clear argument structures, and reader retention strategies. Use when the user asks to write an in-depth article, blog post, newsletter, social media thread, or any content that needs to hold readers' attention. Also use when the user says \"write an article about [topic]\" or \"turn these notes into a publishable piece\". Skill: deep-content-writer Owner: aiepco Summary: Creates engaging, example-driven long-form articles with strong hooks, clear argument structures, and reader retention strategies. Use when the user asks to write an in-depth article, blog post, newsletter, social media thread, or any content that needs to hold readers' attention. Also use when the user says \"write an article about [topic]\" or \"turn these notes into a","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":1807,"uniquenessScore":49,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-09T13:33:54.156Z","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":"2026-10-09T13:33:54.156Z","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":"2026-10-09T17:37:44.071Z","emptyReason":null},"items":[{"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":"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":"2026-04-10T18:48:31.762Z","createdAt":"2026-02-25T03:38:16.584Z","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"}]}}}