[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ASEAY489NMP7136DV9M99S":3,"skill-same:01M7ASEAY489NMP7136DV9M99S":61},{"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":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":30,"commitSha":31,"skillDirectory":32,"forksCount":33,"repositoryHtmlUrl":34,"repositoryHomepageUrl":6,"ownerHtmlUrl":35,"ownerBlogUrl":36,"ownerDescription":6,"updatedAt":37,"install":38,"resources":50},"01M7ASEAY489NMP7136DV9M99S","codex-model-routing",null,"Choose cost-efficient Codex\u002FOpenAI models (Astra, Sol, Terra, Luna), reasoning effort (Light\u002FLow, Medium, High, Extra High\u002Fxhigh, Max, Ultra), and Fast mode. Use when the user asks for model routing, effort selection, consumption optimization, spending analysis, or whether Fast is worthwhile in Codex. Do not activate for ordinary programming without a routing, cost, model, or effort request, or for routing other providers.","saski\u002Farnesto","saski","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F140356?v=4","01M7ASBDSNXTRYDYB4K1PEA1RV","01M7ASBDTAKBDAAV99S1GE7HK6",6,"Unlicense","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcodex-model-routing\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: codex-model-routing\ndescription: Choose cost-efficient Codex\u002FOpenAI models (Astra, Sol, Terra, Luna), reasoning effort (Light\u002FLow, Medium, High, Extra High\u002Fxhigh, Max, Ultra), and Fast mode. Use when the user asks for model routing, effort selection, consumption optimization, spending analysis, or whether Fast is worthwhile in Codex. Do not activate for ordinary programming without a routing, cost, model, or effort request, or for routing other providers.\nmetadata:\n  category: tool-reference\n  pattern: tool-wrapper\n  owner: engineering\n  status: active\n  review_cycle_days: 60\n  benchmark_after_model_update: true\n---\n\n# Codex Model Routing\n\nRecommend the least expensive route likely to produce an accepted result.\nThis is an on-demand advisory policy for Codex\u002FOpenAI only. Preserve existing\nmodel-routing criteria for Claude, Gemini, Hermes, OmniRoute, OpenCode, and\nother providers. Do not change global configuration, model defaults, symlinks,\nhooks, or MCPs as part of a recommendation. A recommendation is not evidence\nthat the active model, effort, or Fast setting changed.\n\n## Choose a route\n\n1. Identify the deliverable, acceptance check, ambiguity, risk, dependencies,\n   and value of reduced latency. Use supplied context; ask only when a missing\n   constraint materially changes the choice.\n2. Select a model from the policy below. Terra Medium is the default\n   recommendation, not a configuration change or an obligation to use Terra\n   for a clearly harder task.\n3. Select effort separately from Fast. Explain the cheapest plausible route\n   and the evidence that would justify escalation.\n4. Return the model, effort, Fast on\u002Foff, brief reason, acceptance check, and\n   stop\u002Fescalation condition. For a spending review, also report measurement\n   gaps and a small comparison plan.\n\n| Work | Starting model and effort |\n|---|---|\n| Clear, repeatable extraction, classification, transformation, or mechanical checks with defined input\u002Foutput | Luna Light\u002FLow; Medium if it needs several bounded steps |\n| Everyday bounded implementation, documents, reporting, or analysis with tools | Terra Medium |\n| Several sources, dependent steps, edge cases, or explicit trade-offs within a clear specification | Terra High |\n| Open-ended diagnosis, architecture, cross-repository complexity, consequential delivery, or deep risk review | Sol Medium or High; High for difficult diagnosis or risk |\n| Exceptionally difficult end-to-end workflows combining code, research, computer use, and sustained judgment | Astra Light\u002FLow or Medium; justify why Sol is insufficient |\n\nDo not infer that a more expensive model is better from spend per turn alone.\nFor the source rationale and measurement limits, read\n[policy evidence](references\u002Fpolicy-evidence.md).\n\n## Effort and continuity guardrails\n\nUse Light\u002FLow (`low`) for a clear transformation, Medium (`medium`) for a\nnormal verifiable unit, and High (`high`) for deeper dependent reasoning.\nExtra High maps to `xhigh`; Max to `max`; Ultra to `ultra`. These are policy\nlabels, not a guarantee every model\u002Fclient supports every combination.\n\n- Never raise effort to repair a session that interrupts or loses continuity.\n  Prepare a concise handoff with the goal, relevant files, verified facts,\n  attempted fixes, remaining failure, and next acceptance check. Start a fresh,\n  bounded unit with that handoff; use `AGENT_HANDOFF.md` when a file is useful.\n- In the Terra recovery path, escalate that unit to Sol High only after the\n  failure repeats in the fresh, clean unit. An initial interruption alone is\n  insufficient. Distinguish model failure from missing access or broken tools;\n  higher reasoning cannot repair those prerequisites. This recovery gate does\n  not prevent selecting Sol initially for clearly complex or risky work.\n- Extra High, Max, and Ultra each require an explicit reason, verifiable\n  acceptance criterion, and checkpoint or stop condition. State why a cheaper\n  route is insufficient; do not recommend them for vague exploration.\n- Extra High is for a difficult, high-value, well-scoped question. Max is for\n  one genuinely indivisible problem where depth outweighs cost and delay.\n- Ultra is only for genuinely independent subproblems with separate verifiable\n  deliverables. Identify the independent boundaries and integration check.\n  Do not use Ultra for one deep problem or tightly dependent steps, and do not\n  launch subagents merely because the user asked for a routing recommendation.\n\n## Fast is a separate spending decision\n\nDefault to Fast off (Standard). Fast reduces latency and increases spending;\nit is not greater reasoning effort and cannot substitute for it. Recommend\nFast only when the value of faster interactive iteration justifies its cost.\nRecord Fast separately from model and effort. Verify the current surcharge\nbefore quoting a number; do not reuse a historical multiplier as a live price.\n\n## Current prices and availability\n\nFor current prices, credit rates, Fast surcharges, model IDs, availability, or\nsupported effort combinations, consult official OpenAI documentation at answer\ntime. Start with the official links in [policy evidence](references\u002Fpolicy-evidence.md),\nfollow official replacements if needed, and state the date and relevant\nplan\u002Fclient limits. Use the active documentation-lookup workflow when available.\nIf verification is unavailable, say so and omit a definitive current quote.\nNever treat the September 2026 documents as a permanent price list.\n\n## Measure accepted outcomes\n\nCompare at least three equivalent tasks per candidate route where practical.\nKeep task scope and acceptance criteria comparable. Record:\n\n| Task\u002Funit | Accepted result | Retries | Human minutes | Credits | Model | Effort | Fast |\n|---|---|---|---|---|---|---|---|\n| Identifier and acceptance check | Yes\u002Fno, with evidence | Count | Review and correction time | Measured or estimated, with rate date | Actual model | Actual effort | On\u002Foff\u002Funknown |\n\nCount failed attempts in total cost. Prefer accepted results per credit and\nper human minute over raw token or message counts. Keep historical consumption,\nproduct limits, estimated equivalent credits, and reconciled billing separate.\nDo not infer Fast from token counts or invent missing acceptance\u002Ftime data.\nRead [evaluation plan](references\u002Feval-plan.md) when testing or revising this skill.\n","# Codex Model Routing\n\nRecommend the least expensive route likely to produce an accepted result.\nThis is an on-demand advisory policy for Codex\u002FOpenAI only. Preserve existing\nmodel-routing criteria for Claude, Gemini, Hermes, OmniRoute, OpenCode, and\nother providers. Do not change global configuration, model defaults, symlinks,\nhooks, or MCPs as part of a recommendation. A recommendation is not evidence\nthat the active model, effort, or Fast setting changed.\n\n## Choose a route\n\n1. Identify the deliverable, acceptance check, ambiguity, risk, dependencies,\n   and value of reduced latency. Use supplied context; ask only when a missing\n   constraint materially changes the choice.\n2. Select a model from the policy below. Terra Medium is the default\n   recommendation, not a configuration change or an obligation to use Terra\n   for a clearly harder task.\n3. Select effort separately from Fast. Explain the cheapest plausible route\n   and the evidence that would justify escalation.\n4. Return the model, effort, Fast on\u002Foff, brief reason, acceptance check, and\n   stop\u002Fescalation condition. For a spending review, also report measurement\n   gaps and a small comparison plan.\n\n| Work | Starting model and effort |\n|---|---|\n| Clear, repeatable extraction, classification, transformation, or mechanical checks with defined input\u002Foutput | Luna Light\u002FLow; Medium if it needs several bounded steps |\n| Everyday bounded implementation, documents, reporting, or analysis with tools | Terra Medium |\n| Several sources, dependent steps, edge cases, or explicit trade-offs within a clear specification | Terra High |\n| Open-ended diagnosis, architecture, cross-repository complexity, consequential delivery, or deep risk review | Sol Medium or High; High for difficult diagnosis or risk |\n| Exceptionally difficult end-to-end workflows combining code, research, computer use, and sustained judgment | Astra Light\u002FLow or Medium; justify why Sol is insufficient |\n\nDo not infer that a more expensive model is better from spend per turn alone.\nFor the source rationale and measurement limits, read\n[policy evidence](references\u002Fpolicy-evidence.md).\n\n## Effort and continuity guardrails\n\nUse Light\u002FLow (`low`) for a clear transformation, Medium (`medium`) for a\nnormal verifiable unit, and High (`high`) for deeper dependent reasoning.\nExtra High maps to `xhigh`; Max to `max`; Ultra to `ultra`. These are policy\nlabels, not a guarantee every model\u002Fclient supports every combination.\n\n- Never raise effort to repair a session that interrupts or loses continuity.\n  Prepare a concise handoff with the goal, relevant files, verified facts,\n  attempted fixes, remaining failure, and next acceptance check. Start a fresh,\n  bounded unit with that handoff; use `AGENT_HANDOFF.md` when a file is useful.\n- In the Terra recovery path, escalate that unit to Sol High only after the\n  failure repeats in the fresh, clean unit. An initial interruption alone is\n  insufficient. Distinguish model failure from missing access or broken tools;\n  higher reasoning cannot repair those prerequisites. This recovery gate does\n  not prevent selecting Sol initially for clearly complex or risky work.\n- Extra High, Max, and Ultra each require an explicit reason, verifiable\n  acceptance criterion, and checkpoint or stop condition. State why a cheaper\n  route is insufficient; do not recommend them for vague exploration.\n- Extra High is for a difficult, high-value, well-scoped question. Max is for\n  one genuinely indivisible problem where depth outweighs cost and delay.\n- Ultra is only for genuinely independent subproblems with separate verifiable\n  deliverables. Identify the independent boundaries and integration check.\n  Do not use Ultra for one deep problem or tightly dependent steps, and do not\n  launch subagents merely because the user asked for a routing recommendation.\n\n## Fast is a separate spending decision\n\nDefault to Fast off (Standard). Fast reduces latency and increases spending;\nit is not greater reasoning effort and cannot substitute for it. Recommend\nFast only when the value of faster interactive iteration justifies its cost.\nRecord Fast separately from model and effort. Verify the current surcharge\nbefore quoting a number; do not reuse a historical multiplier as a live price.\n\n## Current prices and availability\n\nFor current prices, credit rates, Fast surcharges, model IDs, availability, or\nsupported effort combinations, consult official OpenAI documentation at answer\ntime. Start with the official links in [policy evidence](references\u002Fpolicy-evidence.md),\nfollow official replacements if needed, and state the date and relevant\nplan\u002Fclient limits. Use the active documentation-lookup workflow when available.\nIf verification is unavailable, say so and omit a definitive current quote.\nNever treat the September 2026 documents as a permanent price list.\n\n## Measure accepted outcomes\n\nCompare at least three equivalent tasks per candidate route where practical.\nKeep task scope and acceptance criteria comparable. Record:\n\n| Task\u002Funit | Accepted result | Retries | Human minutes | Credits | Model | Effort | Fast |\n|---|---|---|---|---|---|---|---|\n| Identifier and acceptance check | Yes\u002Fno, with evidence | Count | Review and correction time | Measured or estimated, with rate date | Actual model | Actual effort | On\u002Foff\u002Funknown |\n\nCount failed attempts in total cost. Prefer accepted results per credit and\nper human minute over raw token or message counts. Keep historical consumption,\nproduct limits, estimated equivalent credits, and reconciled billing separate.\nDo not infer Fast from token counts or invent missing acceptance\u002Ftime data.\nRead [evaluation plan](references\u002Feval-plan.md) when testing or revising this skill.\n",{"name":5,"metadata":23,"description":7},{"owner":24,"status":25,"pattern":26,"category":27,"review_cycle_days":28,"benchmark_after_model_update":29},"engineering","active","tool-wrapper","tool-reference",60,true,"main","3c24fc2c035029b08c7326774bee3fef8fb95566",".agents\u002Fskills\u002Fcodex-model-routing",0,"https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto","https:\u002F\u002Fgithub.com\u002Fsaski","http:\u002F\u002Fsaski.com","2026-11-12 20:09:16",{"source":8,"skillName":5,"commands":39,"prompt":49,"zipAvailable":29},[40,43,46],{"tool":41,"command":42},"npx","npx skills add saski\u002Farnesto --skill codex-model-routing",{"tool":44,"command":45},"bunx","bunx skills add saski\u002Farnesto --skill codex-model-routing",{"tool":47,"command":48},"pnpm","pnpm dlx skills add saski\u002Farnesto --skill codex-model-routing","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcodex-model-routing\u002FSKILL.md",[51,57],{"path":52,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":55,"sourceHtmlUrl":56},"references\u002Feval-plan.md","REFERENCE","参考文档",3071,"https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcodex-model-routing\u002Freferences\u002Feval-plan.md",{"path":58,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":59,"sourceHtmlUrl":60},"references\u002Fpolicy-evidence.md",2130,"https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fcodex-model-routing\u002Freferences\u002Fpolicy-evidence.md",[62,69,76,83,90,97],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":18},"01M7AT7F5FSHNAMFMF01N0QA1X","visualize","Use \u002Fvisualize for creating compact charts, diagrams, or other visuals inline as part of the conversation itself. It chooses among Markdown tables, Mermaid, inline HTML, and other supported in-transcript forms. Use Canvas for durable artifacts outside the transcript that users may revisit, refine, or share.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fvisualize\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":18},"01M7AT79MHSAR3JDDMP7DGERM9","update-cursor-settings","Modify Cursor\u002FVSCode user settings in settings.json. Use when you want to change editor settings, preferences, configuration, themes, font size, tab size, format on save, auto save, keybindings, or any settings.json values.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fupdate-cursor-settings\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":18},"01M7AT73WF3Y22E0FKYAMCCJES","update-cli-config","View and modify Cursor CLI configuration settings in ~\u002F.cursor\u002Fcli-config.json. Use when the user wants to change CLI settings, configure permissions, switch approval mode, enable vim mode, toggle display options, configure sandbox, or manage any CLI preferences.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fupdate-cli-config\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":18},"01M7AT6F3FBP85AJJ5BJNR8B5W","statusline","Configure a custom status line in the CLI. Use when the user mentions status line, statusline, statusLine, CLI status bar, prompt footer customization, or wants to add session context above the prompt.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fstatusline\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":18},"01M7AT6D24QFS5XRAFN8S5XJ7W","split-to-prs","Split current work into small reviewable PRs. Use when the user asks to split a chat, set of changes, branch, or PR.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fsplit-to-prs\u002FSKILL.md",[],[],{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":18},"01M7AT69FYFN84XZKW452JV5DF","shell","Runs the rest of a \u002Fshell request as a literal shell command. Use only when the user explicitly invokes \u002Fshell and wants the following text executed directly in the terminal.","https:\u002F\u002Fgithub.com\u002Fsaski\u002Farnesto\u002Fblob\u002Fmain\u002F.cursor\u002Fskills-cursor\u002Fshell\u002FSKILL.md",[],[]]