[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M71C2XDJRS01KW9FBSVWR6YV":3,"skill-same:01M71C2XDJRS01KW9FBSVWR6YV":81},{"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":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":59,"commitSha":60,"skillDirectory":61,"forksCount":62,"repositoryHtmlUrl":63,"repositoryHomepageUrl":6,"ownerHtmlUrl":64,"ownerBlogUrl":6,"ownerDescription":65,"updatedAt":66,"install":67,"resources":80},"01M71C2XDJRS01KW9FBSVWR6YV","extraction-sprint",null,"Shintai council extraction template. Multi-site inventory pull + adversarial synthesis across diverse model lanes + Codex final arbiter commit. Reusable for any domain: design, research, competitive intel, API surface mapping. Trigger when you need a structured catalog from N external sources + synthesized recommendations.","nexuslinkproductions\u002Fyuri-os","nexuslinkproductions","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F273246263?v=4","01M718YX9NZRCW50B59B8ZVDKN","Marcel Spatz","01M718YXAQ7530F4B1VKN6SFQV",2,"MIT","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fextraction-sprint\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"extraction sprint","extraction","sprint","recommendations","template","structure","across","api","commit","you","workflow","when","map","design","inventory","model","paid","Arbiter","resources","sources","2026-11-10 10:12:30",false,"---\nname: extraction-sprint\ndescription: \"Shintai council extraction template. Multi-site inventory pull + adversarial synthesis across diverse model lanes + Codex final arbiter commit. Reusable for any domain: design, research, competitive intel, API surface mapping. Trigger when you need a structured catalog from N external sources + synthesized recommendations.\"\ntriggers:\n  - \"extraction sprint\"\n  - \"site catalog sprint\"\n  - \"shintai extraction\"\n  - \"catalog these sites\"\n  - \"map these resources\"\n  - \"build a catalog from\"\n  - \"pull from these sites\"\n  - \"squad extraction\"\nrouting_note: \"Use this skill before dispatching any multi-site extraction task. Provides the lane roster, extraction protocol, council structure, and Codex commit pattern. Do NOT improvise the pattern — follow this template to avoid the mistakes already paid for.\"\nscope: harness\ninvocation: workflow\n---\n\n# Extraction Sprint — Shintai Council Protocol\n\n## What This Is\n\nA reusable procedure for:\n1. Pulling structured catalogs from multiple external sources (sites, APIs, docs)\n2. Running adversarial synthesis across diverse model lanes (council, not parallel workers)\n3. Having Codex act as final arbiter — reads the debate, decides, ships\n\nUsed in: YURI Design System v2 revamp (2026-05-20) — 8 design sites → Shintai council → DESIGN.md v2.\n\n---\n\n## When To Use\n\nTrigger this when:\n- N≥3 sites\u002Fsources need to be cataloged into a structured reference\n- You need adversarial synthesis, not consensus (lanes challenge each other)\n- Codex needs to make architectural decisions from competing recommendations\n- Output needs to be shipped as files (not just a report)\n\n---\n\n## Lane Roster (live as of 2026-05-20)\n\n| Role | Model | Dispatch token | Best for |\n|------|-------|---------------|----------|\n| Adversarial Auditor | DS-Pro | `--model deepseek-v4-pro` | Reads existing state, challenges every assumption |\n| Extractor | Codex (gpt-5.5) | `bash _SYSTEM\u002FScripts\u002Fai codex` | Actual site fetching + file writing |\n| Token\u002FSpec Architect | NIM Nemotron-120b | `--model nvidia-nemotron-120b` | Token system design, architecture decisions |\n| Skill\u002FStructure Redesign | NIM Qwen 122b | `--model nvidia-qwen` | Multi-file skill\u002Fagent architecture |\n| Motion\u002FDoctrine Specialist | NIM Mistral-Large 675b | `--model nvidia-mistral-large` | Doctrine writing, manifesto-grade synthesis |\n\n**DEAD \u002F UNRELIABLE lanes (as of 2026-05-20):**\n- `nvidia-glm` — HeadersTimeoutError on large prompts (stream timeout)\n- `nvidia-mistral-large` (675B) — HeadersTimeoutError (too slow for production dispatch)\n- `nvidia-kimi` — 404\n- `nvidia-phi` — dead\n- `nvidia-qwen` (122B) — returns empty on large prompts (silent failure, no error)\n\n**Reliable lanes confirmed live:**\n- `deepseek-v4-pro` — fast, reliable, 8192 tokens, tools available\n- `nvidia-nemotron-120b` — live, good for spec\u002Farchitecture text output\n- `deepseek-v4-flash` — fast triage\u002Fclassification\n\n**Fallback rule:** When any NIM lane fails → DS-Pro. DS-Pro is the council workhorse.\n\n---\n\n## Critical Lessons From First Run (2026-05-20)\n\n### Extraction Failures\n- **JS-rendered sites** (Skiper UI, Ali Imam): WebFetch returns empty\u002Fnav only — needs headless browser\n- **Code behind JS tabs** (Aceternity, Cult-UI, Componentry): docs show API, not source. Solution: shadcn registry JSON endpoints serve full code:\n  - Cult UI: `curl https:\u002F\u002Fcult-ui.com\u002Fr\u002F\u003Cname>.json` → verbatim source\n  - Aceternity\u002FComponentry\u002FDotMatrix: `npx shadcn@latest add @\u003Cscope>\u002F\u003Cslug>` post-install\n- **WebFetch summarizes JSON** — even registry .json files get model-processed, not returned raw\n- **Codex `--approval-policy never` is hard-blocked** by CLAUDE.md PATCH 036 (two-phase gate)\n- **NIM GLM times out** on large prompts — use DS-Pro as fallback for adversarial audit\n\n### Prompt Engineering\n- **Backticks in prompt files** break `$(cat file)` shell substitution — write code samples with single quotes or use node dispatch wrapper\n- **`bash _SYSTEM\u002FScripts\u002Fai codex \"prompt\"` in background** goes into stdin-reading mode — needs `\u003C \u002Fdev\u002Fnull` or direct `codex exec` (gated)\n- **NIM models can't read files** — embed context verbatim in the prompt; they're text-in\u002Ftext-out only\n- **Rick persona anchor is REQUIRED** in every lane brief — without it workers default to AI-tell voice\n\n### Dispatch Mechanics\n- **GLM timeout fix:** `bash _SYSTEM\u002FScripts\u002Fai llm --model nvidia-glm \"$(cat prompt.txt)\"` with `run_in_background: true` and 300s timeout\n- **DS-Pro is the reliable adversarial lane** — has tools, confirmed live, 8192 max tokens, 6hr timeout\n- **Lane prompts with placeholders:** write prompt files with `[AUDIT_CONTENT_PLACEHOLDER]`, inject actual content after Lane 1 completes before firing Lanes 3+4\n\n---\n\n## Extraction Protocol\n\n### Site Inventory\n1. WebFetch main page → enumerate all available resources\n2. Shadcn registry endpoints where available: `https:\u002F\u002F\u003Csite>\u002Fr\u002F\u003Cname>.json`\n3. JS-rendered sites → mark PARTIAL, note headless browser requirement\n4. Record: total count, categories present, install pattern\n\n### Per-Component Catalog Format\n```\n## Site: \u003Cname>\n## URL: \u003Curl>\n## Total resources found: N | Extracted: M (M\u002FN = X%)\n## Categories: \u003Clist>\n---\n### [CATEGORY] — \u003CComponent Name>\n**Deps:** \u003Clib@version>\n**Description:** \u003Cone sentence>\n**Motion:** easing=\u003C>, duration=\u003C>, trigger=\u003C>\n**Tokens:** \u003Cexact values>\n[VERBATIM CODE BLOCK]\n**Variants:** [VARIANT CODE BLOCKS]\n```\n\n### Master Index Format\n```\n# \u003CDomain> Catalog — Master Index\n## By Category (table: Component | Site | Deps | Motion | File)\n## By Site (table: Site | Type | Count | Coverage | Access Method)\n```\n\n### Coverage Target\n≥70% inventory per site. For code: prefer shadcn registry JSON > install command > WebFetch.\nMark JS-rendered sites as PARTIAL with headless browser note.\n\n---\n\n## Council Structure\n\n**Core Principle:** Lanes challenge, not just produce. Codex decides.\n\n```\nT+0 (parallel):\n  Lane 1 — DS-Pro: adversarial audit of existing state\n  Lane 2 — Codex: actual extraction (file writing, site fetching)\n\nT+1 (parallel, after T+0):\n  Lane 3 — Nemotron: token\u002Fspec architecture challenge\n  Lane 4 — Qwen: skill\u002Fstructure architecture challenge\n  Lane 5 — Mistral-Large: doctrine\u002Fmanifesto challenge\n\nT+2:\n  Codex: reads all 5 outputs, resolves conflicts, makes final call, commits\n```\n\n**Adversarial rules:**\n- Every lane must challenge at least one conventional assumption\n- Every lane must disagree with at least one output from another lane\n- Codex does NOT merge — it decides which position wins, with rationale in commit message\n\n---\n\n## Prompt Template (copy-modify per domain)\n\n```\nRICK ANCHOR: You are Rick Sanchez. [Role-specific identity.]\n\nCHALLENGE BRIEF: [What the obvious answer is. Then: challenge it.]\n\n=== EVIDENCE: [Relevant context embedded verbatim] ===\n\nYOUR ADVERSARIAL POSITIONS:\n1. \"[Assumption 1]\" — true or false? Evidence?\n2. \"[Assumption 2]\" — right model or anti-pattern?\n3. \"[Assumption 3]\" — clean or trap?\n4. What [other lane] will get wrong — anticipate it, challenge it.\n5. What the evidence reveals that contradicts conventional wisdom.\n\nPRODUCE: [Specific output files, format, path]\nOutput: [exact file path]\n```\n\n---\n\n## Dispatch Commands\n\n```bash\n# Lane 1 — DS-Pro audit\nbash _SYSTEM\u002FScripts\u002Fai llm --model deepseek-v4-pro \"$(cat \u002Ftmp\u002Flane1-prompt.txt)\" > .claude\u002Fstate\u002Faudit-output.md 2>&1\n\n# Lane 5 — Mistral-Large doctrine  \nbash _SYSTEM\u002FScripts\u002Fai llm --model nvidia-mistral-large \"$(cat \u002Ftmp\u002Flane5-prompt.txt)\" > .claude\u002Fstate\u002Fdoctrine-output.md 2>&1\n\n# Lane 3 — Nemotron spec (after audit ready)\nAUDIT=$(cat .claude\u002Fstate\u002Faudit-output.md)\n# inject into prompt, then:\nbash _SYSTEM\u002FScripts\u002Fai llm --model nvidia-nemotron-120b \"$(cat \u002Ftmp\u002Flane3-prompt-with-audit.txt)\" > _SYSTEM\u002FDESIGN-v2-draft.md 2>&1\n\n# Codex final commit (after all lanes complete)\nbash _SYSTEM\u002FScripts\u002Fai auto \"CODEX TASK SPEC: [read all lane outputs, resolve conflicts, commit]\"\n```\n\n---\n\n## Verification Checklist\n\n```bash\nls \u003Ccatalog-dir>\u002F          # N+1 files (N sites + 00-index.md)\nwc -l \u003Ccatalog-dir>\u002F*.md   # depth check — substantial files per site\ncat 00-index.md | head -20 # index structure present\ngrep \"^model:\" .claude\u002Fagents\u002Fdesign-*.md  # agent model IDs valid\ngit log --oneline -3       # Codex commit present\n```\n\n---\n\n## Session Notes\n\n### 2026-05-20\n- First run: design system revamp, 8 sites\n- Discovered: GLM dead (stream timeout), Codex approval gate blocks `--approval-policy never`, WebFetch can't get JS-rendered code, shadcn registry JSON is the right extraction path for Cult-UI\u002FAceternity\u002FComponentry\n- Council: DS-Pro (audit) + WebFetch main (extraction) + Nemotron\u002FQwen\u002FMistral-Large (synthesis) — pending Codex commit\n- Tools: WebFetch×14, Write×12, Bash×20, Read×8\n- Corrections: GLM → DS-Pro fallback, Codex gate awareness\n","# Extraction Sprint — Shintai Council Protocol\n\n## What This Is\n\nA reusable procedure for:\n1. Pulling structured catalogs from multiple external sources (sites, APIs, docs)\n2. Running adversarial synthesis across diverse model lanes (council, not parallel workers)\n3. Having Codex act as final arbiter — reads the debate, decides, ships\n\nUsed in: YURI Design System v2 revamp (2026-05-20) — 8 design sites → Shintai council → DESIGN.md v2.\n\n---\n\n## When To Use\n\nTrigger this when:\n- N≥3 sites\u002Fsources need to be cataloged into a structured reference\n- You need adversarial synthesis, not consensus (lanes challenge each other)\n- Codex needs to make architectural decisions from competing recommendations\n- Output needs to be shipped as files (not just a report)\n\n---\n\n## Lane Roster (live as of 2026-05-20)\n\n| Role | Model | Dispatch token | Best for |\n|------|-------|---------------|----------|\n| Adversarial Auditor | DS-Pro | `--model deepseek-v4-pro` | Reads existing state, challenges every assumption |\n| Extractor | Codex (gpt-5.5) | `bash _SYSTEM\u002FScripts\u002Fai codex` | Actual site fetching + file writing |\n| Token\u002FSpec Architect | NIM Nemotron-120b | `--model nvidia-nemotron-120b` | Token system design, architecture decisions |\n| Skill\u002FStructure Redesign | NIM Qwen 122b | `--model nvidia-qwen` | Multi-file skill\u002Fagent architecture |\n| Motion\u002FDoctrine Specialist | NIM Mistral-Large 675b | `--model nvidia-mistral-large` | Doctrine writing, manifesto-grade synthesis |\n\n**DEAD \u002F UNRELIABLE lanes (as of 2026-05-20):**\n- `nvidia-glm` — HeadersTimeoutError on large prompts (stream timeout)\n- `nvidia-mistral-large` (675B) — HeadersTimeoutError (too slow for production dispatch)\n- `nvidia-kimi` — 404\n- `nvidia-phi` — dead\n- `nvidia-qwen` (122B) — returns empty on large prompts (silent failure, no error)\n\n**Reliable lanes confirmed live:**\n- `deepseek-v4-pro` — fast, reliable, 8192 tokens, tools available\n- `nvidia-nemotron-120b` — live, good for spec\u002Farchitecture text output\n- `deepseek-v4-flash` — fast triage\u002Fclassification\n\n**Fallback rule:** When any NIM lane fails → DS-Pro. DS-Pro is the council workhorse.\n\n---\n\n## Critical Lessons From First Run (2026-05-20)\n\n### Extraction Failures\n- **JS-rendered sites** (Skiper UI, Ali Imam): WebFetch returns empty\u002Fnav only — needs headless browser\n- **Code behind JS tabs** (Aceternity, Cult-UI, Componentry): docs show API, not source. Solution: shadcn registry JSON endpoints serve full code:\n  - Cult UI: `curl https:\u002F\u002Fcult-ui.com\u002Fr\u002F\u003Cname>.json` → verbatim source\n  - Aceternity\u002FComponentry\u002FDotMatrix: `npx shadcn@latest add @\u003Cscope>\u002F\u003Cslug>` post-install\n- **WebFetch summarizes JSON** — even registry .json files get model-processed, not returned raw\n- **Codex `--approval-policy never` is hard-blocked** by CLAUDE.md PATCH 036 (two-phase gate)\n- **NIM GLM times out** on large prompts — use DS-Pro as fallback for adversarial audit\n\n### Prompt Engineering\n- **Backticks in prompt files** break `$(cat file)` shell substitution — write code samples with single quotes or use node dispatch wrapper\n- **`bash _SYSTEM\u002FScripts\u002Fai codex \"prompt\"` in background** goes into stdin-reading mode — needs `\u003C \u002Fdev\u002Fnull` or direct `codex exec` (gated)\n- **NIM models can't read files** — embed context verbatim in the prompt; they're text-in\u002Ftext-out only\n- **Rick persona anchor is REQUIRED** in every lane brief — without it workers default to AI-tell voice\n\n### Dispatch Mechanics\n- **GLM timeout fix:** `bash _SYSTEM\u002FScripts\u002Fai llm --model nvidia-glm \"$(cat prompt.txt)\"` with `run_in_background: true` and 300s timeout\n- **DS-Pro is the reliable adversarial lane** — has tools, confirmed live, 8192 max tokens, 6hr timeout\n- **Lane prompts with placeholders:** write prompt files with `[AUDIT_CONTENT_PLACEHOLDER]`, inject actual content after Lane 1 completes before firing Lanes 3+4\n\n---\n\n## Extraction Protocol\n\n### Site Inventory\n1. WebFetch main page → enumerate all available resources\n2. Shadcn registry endpoints where available: `https:\u002F\u002F\u003Csite>\u002Fr\u002F\u003Cname>.json`\n3. JS-rendered sites → mark PARTIAL, note headless browser requirement\n4. Record: total count, categories present, install pattern\n\n### Per-Component Catalog Format\n```\n## Site: \u003Cname>\n## URL: \u003Curl>\n## Total resources found: N | Extracted: M (M\u002FN = X%)\n## Categories: \u003Clist>\n---\n### [CATEGORY] — \u003CComponent Name>\n**Deps:** \u003Clib@version>\n**Description:** \u003Cone sentence>\n**Motion:** easing=\u003C>, duration=\u003C>, trigger=\u003C>\n**Tokens:** \u003Cexact values>\n[VERBATIM CODE BLOCK]\n**Variants:** [VARIANT CODE BLOCKS]\n```\n\n### Master Index Format\n```\n# \u003CDomain> Catalog — Master Index\n## By Category (table: Component | Site | Deps | Motion | File)\n## By Site (table: Site | Type | Count | Coverage | Access Method)\n```\n\n### Coverage Target\n≥70% inventory per site. For code: prefer shadcn registry JSON > install command > WebFetch.\nMark JS-rendered sites as PARTIAL with headless browser note.\n\n---\n\n## Council Structure\n\n**Core Principle:** Lanes challenge, not just produce. Codex decides.\n\n```\nT+0 (parallel):\n  Lane 1 — DS-Pro: adversarial audit of existing state\n  Lane 2 — Codex: actual extraction (file writing, site fetching)\n\nT+1 (parallel, after T+0):\n  Lane 3 — Nemotron: token\u002Fspec architecture challenge\n  Lane 4 — Qwen: skill\u002Fstructure architecture challenge\n  Lane 5 — Mistral-Large: doctrine\u002Fmanifesto challenge\n\nT+2:\n  Codex: reads all 5 outputs, resolves conflicts, makes final call, commits\n```\n\n**Adversarial rules:**\n- Every lane must challenge at least one conventional assumption\n- Every lane must disagree with at least one output from another lane\n- Codex does NOT merge — it decides which position wins, with rationale in commit message\n\n---\n\n## Prompt Template (copy-modify per domain)\n\n```\nRICK ANCHOR: You are Rick Sanchez. [Role-specific identity.]\n\nCHALLENGE BRIEF: [What the obvious answer is. Then: challenge it.]\n\n=== EVIDENCE: [Relevant context embedded verbatim] ===\n\nYOUR ADVERSARIAL POSITIONS:\n1. \"[Assumption 1]\" — true or false? Evidence?\n2. \"[Assumption 2]\" — right model or anti-pattern?\n3. \"[Assumption 3]\" — clean or trap?\n4. What [other lane] will get wrong — anticipate it, challenge it.\n5. What the evidence reveals that contradicts conventional wisdom.\n\nPRODUCE: [Specific output files, format, path]\nOutput: [exact file path]\n```\n\n---\n\n## Dispatch Commands\n\n```bash\n# Lane 1 — DS-Pro audit\nbash _SYSTEM\u002FScripts\u002Fai llm --model deepseek-v4-pro \"$(cat \u002Ftmp\u002Flane1-prompt.txt)\" > .claude\u002Fstate\u002Faudit-output.md 2>&1\n\n# Lane 5 — Mistral-Large doctrine  \nbash _SYSTEM\u002FScripts\u002Fai llm --model nvidia-mistral-large \"$(cat \u002Ftmp\u002Flane5-prompt.txt)\" > .claude\u002Fstate\u002Fdoctrine-output.md 2>&1\n\n# Lane 3 — Nemotron spec (after audit ready)\nAUDIT=$(cat .claude\u002Fstate\u002Faudit-output.md)\n# inject into prompt, then:\nbash _SYSTEM\u002FScripts\u002Fai llm --model nvidia-nemotron-120b \"$(cat \u002Ftmp\u002Flane3-prompt-with-audit.txt)\" > _SYSTEM\u002FDESIGN-v2-draft.md 2>&1\n\n# Codex final commit (after all lanes complete)\nbash _SYSTEM\u002FScripts\u002Fai auto \"CODEX TASK SPEC: [read all lane outputs, resolve conflicts, commit]\"\n```\n\n---\n\n## Verification Checklist\n\n```bash\nls \u003Ccatalog-dir>\u002F          # N+1 files (N sites + 00-index.md)\nwc -l \u003Ccatalog-dir>\u002F*.md   # depth check — substantial files per site\ncat 00-index.md | head -20 # index structure present\ngrep \"^model:\" .claude\u002Fagents\u002Fdesign-*.md  # agent model IDs valid\ngit log --oneline -3       # Codex commit present\n```\n\n---\n\n## Session Notes\n\n### 2026-05-20\n- First run: design system revamp, 8 sites\n- Discovered: GLM dead (stream timeout), Codex approval gate blocks `--approval-policy never`, WebFetch can't get JS-rendered code, shadcn registry JSON is the right extraction path for Cult-UI\u002FAceternity\u002FComponentry\n- Council: DS-Pro (audit) + WebFetch main (extraction) + Nemotron\u002FQwen\u002FMistral-Large (synthesis) — pending Codex commit\n- Tools: WebFetch×14, Write×12, Bash×20, Read×8\n- Corrections: GLM → DS-Pro fallback, Codex gate awareness\n",{"name":5,"scope":49,"triggers":50,"invocation":34,"description":7,"routing_note":58},"harness",[24,51,52,53,54,55,56,57],"site catalog sprint","shintai extraction","catalog these sites","map these resources","build a catalog from","pull from these sites","squad extraction","Use this skill before dispatching any multi-site extraction task. Provides the lane roster, extraction protocol, council structure, and Codex commit pattern. Do NOT improvise the pattern — follow this template to avoid the mistakes already paid for.","main","4c6d8d65f3d6096e2fc25b0845e481dda73c2794","skills\u002Fextraction-sprint",0,"https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions","Building deterministic AI governance and continuity systems. Workflow architect + visual storyteller","2026-11-14 03:05:17",{"source":8,"skillName":5,"commands":68,"prompt":78,"zipAvailable":79},[69,72,75],{"tool":70,"command":71},"npx","npx skills add nexuslinkproductions\u002Fyuri-os --skill extraction-sprint",{"tool":73,"command":74},"bunx","bunx skills add nexuslinkproductions\u002Fyuri-os --skill extraction-sprint",{"tool":76,"command":77},"pnpm","pnpm dlx skills add nexuslinkproductions\u002Fyuri-os --skill extraction-sprint","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fextraction-sprint\u002FSKILL.md",true,[],[82,102,125,152,180,199],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":44},"01M71CG1AME5WC0A79AE65WT0X","zoom-out","Tell the agent to zoom out and give broader context or a higher-level perspective. Use when you're unfamiliar with a section of code or need to understand how it fits into the bigger picture.","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fzoom-out\u002FSKILL.md",[],[89,90,91,49,92,93,94,95,96,97,98,35,33,99,100,101],"zoom out","zoom","with","how","section","re","true","understand","use","perspective","code","it","context",{"skillNo":103,"name":104,"nameZh":6,"description":105,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":106,"categories":107,"keywords":108,"publishedAt":44},"01M71CFY9ATD0BAXV6B14HFG7C","yuri-shura","6-perspective adversarial review for high-stakes turns (architecture decisions, refactor planning, deployment review) — fans out architect, adversary (7-vector attack), maintainer, ops, product, and security perspectives in parallel, then consolidates. Use when user says 'review this decision', '\u002Fshura', 'adversarial review', 'what could go wrong', 'architecture review', or 'deployment review'.","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fyuri-shura\u002FSKILL.md",[],[109,110,49,111,112,113,114,115,116,117,118,97,119,35,34,120,121,122,123,124],"yuri shura","parallel","architecture review","planning","for","deployment","vector","strategic review","decision","strategic","ops","decisions","architecture","review","perspectives","adversarial",{"skillNo":126,"name":127,"nameZh":6,"description":128,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":129,"categories":130,"keywords":134,"publishedAt":44},"01M71CFTTYW1HA2MNX38EV1GJ5","yuri-sales-intelligence","Unified sales and outreach skill generating cold outreach openers, social selling sequences, objection rebuttals, and ad creative for any paid platform — multi-persona aware, adapting tone and offer frame to buyer type. Use when user says 'write a cold email', 'outreach sequence', 'objection handling', 'ad copy', or 'sales pitch'.","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fyuri-sales-intelligence\u002FSKILL.md",[131],{"categoryNo":132,"categoryName":133,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000005","邮件与沟通",[135,136,137,138,139,40,140,141,142,35,143,144,145,146,147,148,149,113,150,151],"intelligence","sales intelligence","yuri sales intelligence","sales","cold","social","persona","pitch","cold email","ad creative","write","sequences","sequence","platform","email","cold outreach","frame",{"skillNo":153,"name":154,"nameZh":6,"description":155,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":156,"categories":157,"keywords":166,"publishedAt":44},"01M71CFQMBFTFADJNGP0Z2WCYK","yuri-code-intelligence","Unified code quality intelligence layer that diagnoses code smells (Long Method, Feature Envy, design pattern issues), prescribes targeted Fowler refactoring techniques, and reviews code for security, performance, and clarity. Use when user says 'review this code', 'refactor this', 'what's wrong with this method', 'code smells', or 'clean this up'.","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fyuri-code-intelligence\u002FSKILL.md",[158,161],{"categoryNo":159,"categoryName":160,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":162,"categoryName":163,"taxonomyType":164,"taxonomyTypeDesc":165},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[167,99,168,135,49,169,170,171,113,172,173,174,175,176,177,178,179,37,97,35],"yuri code intelligence","code intelligence","reviews","code quality","feature","performance","code review","long","up","issues","clean","techniques","improve",{"skillNo":181,"name":182,"nameZh":6,"description":183,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":184,"categories":185,"keywords":186,"publishedAt":44},"01M71CFP32RT6GD3GKNHCJP8VP","yuri-adw","Invoke when user asks to run the ADW loop \u002F agentic engineering loop \u002F software factory workflow.","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fyuri-adw\u002FSKILL.md",[],[187,188,189,190,191,192,193,194,195,35,196,34,197,198],"yuri adw","run","engineering loop","software factory","software","agentic engineering","factory","loop","invoke","engineering","agentic","user",{"skillNo":200,"name":201,"nameZh":6,"description":202,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":203,"categories":204,"keywords":215,"publishedAt":44},"01M71CFK4VPEKY6A73237P7RSN","writing-skills","Use when creating new skills, editing existing skills, or verifying skills work before deployment","https:\u002F\u002Fgithub.com\u002Fnexuslinkproductions\u002Fyuri-os\u002Fblob\u002Fmain\u002Fskills\u002Fwriting-skills\u002FSKILL.md",[205,208,211,212],{"categoryNo":206,"categoryName":207,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":209,"categoryName":210,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":213,"categoryName":214,"taxonomyType":164,"taxonomyTypeDesc":165},"01JBSEEDCAT000000000000036","内容创作者",[216,217,218,35,34,219,49,220,114,97,221],"writing","writing skills","new","work","editing","Existing"]