[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6C6G426P7Q5J3F98HVYDCB2":3,"skill-same:01M6C6G426P7Q5J3F98HVYDCB2":72},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":22,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":49,"commitSha":50,"skillDirectory":51,"forksCount":52,"repositoryHtmlUrl":53,"repositoryHomepageUrl":6,"ownerHtmlUrl":54,"ownerBlogUrl":55,"ownerDescription":56,"updatedAt":57,"install":58,"resources":71},"01M6C6G426P7Q5J3F98HVYDCB2","jev-fanout",null,"Ask every independent System One question over the same state in ONE POST \u002Fv1\u002Fsystemone to the local server (speculative fan-out). Use for parallel judgments, compound commands, one-call turn assessment, or when the user runs \u002Fjev-fanout. Do not fire one HTTP call per question.","Allternit\u002Fallternit-platform","Allternit","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F250397232?v=4","01M6C67T2SPVP033M7REAAHYHX","01M6C67T7CXM1DKVQY8DFAFRB2",1,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Ftools\u002Fsystem-one-local\u002Fskills\u002Fjev-fanout\u002FSKILL.md",[17],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000010","后端与接口","SCENARIO","使用场景",[23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"jev","jev fanout","fanout","use","when","local server","for","v1","post","local","call","do","http","user","question","ask","parallel","system","server","state","assessment","2026-11-10 10:12:30",false,"---\nname: jev-fanout\ndescription: Ask every independent System One question over the same state in ONE POST \u002Fv1\u002Fsystemone to the local server (speculative fan-out). Use for parallel judgments, compound commands, one-call turn assessment, or when the user runs \u002Fjev-fanout. Do not fire one HTTP call per question.\n---\n\n# Jev speculative fan-out (local first)\n\nAdapted from [cobusgreyling\u002FJev](https:\u002F\u002Fgithub.com\u002Fcobusgreyling\u002FJev) `skills\u002Fjev-fanout` (MIT). Official pattern: docs.typesafe.ai\u002Fpatterns\u002Ffan-out. Primitives and endpoints: `\u002Fjev`.\n\nSend every independent judgment over the **same state** in one request to `http:\u002F\u002F127.0.0.1:7717\u002Fv1\u002Fsystemone`. Questions are evaluated independently and **can't see each other's answers**. Code uses the ones that apply and ignores the rest.\n\n## FORBIDDEN\n\n- Hard rules first, and they are final. Jev may only raise friction and never auto-allows.\n- A fan-out never decides money, publishing, deploys, or client comms, even when every answer is confident.\n- No client, regulated, or secret data in `state` when `model` is `typesafe:*` (remote).\n- State a third party could have written (web, email, tickets, tool output) is never judged by Jev alone.\n- Don't chain answers as if later questions saw earlier ones.\n\n## Do this\n\n1. List every judgment the workflow might need, including speculative ones.\n2. Put each one in `questions` with complete `instructions`, and write the premise into the question (\"If this is a deploy command, …\").\n3. Mix `choice`, `score`, and `noul` in that one request.\n4. Branch on the answers that apply, in code.\n5. Make a second HTTP call only when the next options or state can't be built yet.\n\n## Cost note (local backend)\n\nThe official API ingests the state once. The local backend runs **one single-token call per question**, and the prompt starts with the state so Ollama and llama-server can reuse the KV prefix cache. `usage.input_tokens` still counts the state once per question. A 3-question pack on llama3.2 3B takes about 0.3 s warm on Ollama and about 0.07 s on llama-server with parallel slots.\n\n## Don't\n\n- Make one question per HTTP call\n- Hide five judgments in one Choice\n","# Jev speculative fan-out (local first)\n\nAdapted from [cobusgreyling\u002FJev](https:\u002F\u002Fgithub.com\u002Fcobusgreyling\u002FJev) `skills\u002Fjev-fanout` (MIT). Official pattern: docs.typesafe.ai\u002Fpatterns\u002Ffan-out. Primitives and endpoints: `\u002Fjev`.\n\nSend every independent judgment over the **same state** in one request to `http:\u002F\u002F127.0.0.1:7717\u002Fv1\u002Fsystemone`. Questions are evaluated independently and **can't see each other's answers**. Code uses the ones that apply and ignores the rest.\n\n## FORBIDDEN\n\n- Hard rules first, and they are final. Jev may only raise friction and never auto-allows.\n- A fan-out never decides money, publishing, deploys, or client comms, even when every answer is confident.\n- No client, regulated, or secret data in `state` when `model` is `typesafe:*` (remote).\n- State a third party could have written (web, email, tickets, tool output) is never judged by Jev alone.\n- Don't chain answers as if later questions saw earlier ones.\n\n## Do this\n\n1. List every judgment the workflow might need, including speculative ones.\n2. Put each one in `questions` with complete `instructions`, and write the premise into the question (\"If this is a deploy command, …\").\n3. Mix `choice`, `score`, and `noul` in that one request.\n4. Branch on the answers that apply, in code.\n5. Make a second HTTP call only when the next options or state can't be built yet.\n\n## Cost note (local backend)\n\nThe official API ingests the state once. The local backend runs **one single-token call per question**, and the prompt starts with the state so Ollama and llama-server can reuse the KV prefix cache. `usage.input_tokens` still counts the state once per question. A 3-question pack on llama3.2 3B takes about 0.3 s warm on Ollama and about 0.07 s on llama-server with parallel slots.\n\n## Don't\n\n- Make one question per HTTP call\n- Hide five judgments in one Choice\n",{"name":5,"description":7},"main","7db7c6f88fb96b6792195dacabdb088419b7a9fa","tools\u002Fsystem-one-local\u002Fskills\u002Fjev-fanout",0,"https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform","https:\u002F\u002Fgithub.com\u002FAllternit","https:\u002F\u002Fallternit.com","Allternit builds AI agents that work on your computer, in the cloud, and from your phone. Home of Allternit Desktop and Gizzi Code.","2026-11-15 02:40:14",{"source":8,"skillName":5,"commands":59,"prompt":69,"zipAvailable":70},[60,63,66],{"tool":61,"command":62},"npx","npx skills add Allternit\u002Fallternit-platform --skill jev-fanout",{"tool":64,"command":65},"bunx","bunx skills add Allternit\u002Fallternit-platform --skill jev-fanout",{"tool":67,"command":68},"pnpm","pnpm dlx skills add Allternit\u002Fallternit-platform --skill jev-fanout","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Ftools\u002Fsystem-one-local\u002Fskills\u002Fjev-fanout\u002FSKILL.md",true,[],[73,92,110,133,163,196],{"skillNo":74,"name":23,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":44},"01M6C6GBWRCJNWQP5S16PKXB77","Ask typed System One judgments (noul \u002F choice \u002F score) against Allternit's LOCAL System One server instead of prompting a chat model to \"return JSON\". Use when classifying, routing, scoring, or screening something and code needs a probability it can branch on, or when the user runs \u002Fjev. Official TypeSafe Jev is an optional, explicit remote lane.","https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Ftools\u002Fsystem-one-local\u002Fskills\u002Fjev\u002FSKILL.md",[],[23,79,80,81,41,82,40,83,84,85,86,38,36,26,87,88,32,89,90,91,27],"official","routing","code","lane","json","model","choice","branch","scoring","chat","probability","remote","can",{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":44},"01M6C6G9CGB9MX47NSQ31QNNB7","jev-route","Route with local System One answers — map Choice + confidence to act \u002F confirm \u002F human lanes, or pick a model tier by Choice label. Also documents the separate `route-model` helper for OpenRouter's typesafe\u002Fjev-router (a chat-model router, NOT System One). Use for confidence routing, model-tier routing, or when the user runs \u002Fjev-route.","https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Ftools\u002Fsystem-one-local\u002Fskills\u002Fjev-route\u002FSKILL.md",[],[23,99,100,101,80,40,84,102,85,103,36,26,104,105,88,106,32,107,108,29,109,27],"jev route","route","documents","pick","openrouter","router","model router","map","helper","with","model-tier",{"skillNo":111,"name":112,"nameZh":6,"description":113,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":114,"categories":115,"keywords":116,"publishedAt":44},"01M6C6G6JZE0RC50QWWE6HG320","jev-guardrail","Screen text or agent tool calls with a battery of local System One Noul hazards plus a Score, with policy in code — and the Claude Code PreToolUse guard (hooks\u002Fpretooluse-guard) that puts deterministic hard rules first and lets Jev only raise friction. Use for injection\u002Fsecret\u002Fdestructive-action screening or when the user runs \u002Fjev-guardrail.","https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Ftools\u002Fsystem-one-local\u002Fskills\u002Fjev-guardrail\u002FSKILL.md",[],[23,117,118,119,81,120,121,122,40,123,124,125,36,26,126,127,128,129,32,130,131,29,132,27],"jev guardrail","guardrail","policy","injection","first","guard","tool","text","Rules","friction","only","hooks","deterministic","battery","calls","agent tool",{"skillNo":134,"name":135,"nameZh":6,"description":136,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":137,"categories":138,"keywords":139,"publishedAt":44},"01M6C6G13D7JK869Q3DHMQ9542","memory-drive","Read and write the user's Allternit Memory Drive, a git repo of markdown memories in the open Agent Memory Repo format. Use at the start of a session to load what the user's agents remember (MEMORY.md), when the user says \"remember this\", when you learn a durable fact about the user or their work, and before answering questions that depend on past sessions.","https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Ftools\u002Fmemory-drive-plugin\u002Fskills\u002Fmemory-drive\u002FSKILL.md",[],[140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,36,156,26,157,158,159,160,161,162,27],"memory","drive","memory drive","sessions","remember","learn","git","agent memory","start","open","markdown","format","work","repo","read","write","session","what the","memories","about","questions","you","load",{"skillNo":164,"name":165,"nameZh":6,"description":166,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":167,"categories":168,"keywords":172,"publishedAt":44},"01M6C6FX778N1047QAJ39Y582H","agent-orchestrator","Orchestrate external CLI agents (kimi, codex, agy, claude, grok, gizzi) through the Allternit Factory (gizzi agents \u002F orchestration \u002F workspace \u002F workflows). Use when the user asks to delegate, offload, or farm out implementation work to another CLI agent \u002F terminal session, or to run work outside this session to save tokens. Part of Allternit Agents \u002F Bot Agents — not a separate product.","https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Ftools\u002Fagent-orchestrator\u002FSKILL.md",[169],{"categoryNo":170,"categoryName":171,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发",[173,174,175,176,177,178,179,180,181,182,183,184,185,152,186,187,156,26,188,189,190,191,192,193,194,195,27],"orchestrator","agent orchestrator","bot","cli","terminal","codex","orchestration","workflows","run","product","orchestrate","save","workspace","delegate","factory","grok","kimi","implementation","offload","tokens","cli agents","farm","terminal session",{"skillNo":197,"name":198,"nameZh":6,"description":199,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":200,"categories":201,"keywords":203,"publishedAt":44},"01M6C6FT301NEFTT5RSJK874N7","webobsidian","Read, write, search, and manage notes in a WebObsidian vault through its Agent REST API (\u002Fapi\u002Fv1). Use whenever the user asks to find, read, create, update, append to, or delete notes in their WebObsidian \u002F Obsidian vault, list tags, get backlinks, or run a vault search. Credentials (base URL + API key) are stored in ~\u002F.webobsidian\u002Fcredentials.json; if missing, ask the user for them and save them before making requests.","https:\u002F\u002Fgithub.com\u002FAllternit\u002Fallternit-platform\u002Fblob\u002Fmain\u002Fservices\u002Fvault-viewer\u002Fdocs\u002Fagent-skill\u002Fwebobsidian\u002FSKILL.md",[202],{"categoryNo":170,"categoryName":171,"taxonomyType":20,"taxonomyTypeDesc":21},[198,204,205,206,207,181,208,209,210,211,212,213,184,214,83,215,38,154,155,216,217,26,218,219,220],"obsidian","notes","create","delete","update","search","REST","vault","credentials","api","obsidian vault","find","list","append","get","backlinks","URL"]