[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7B2K2PBDHCS5SDJHHPEV2ZW":3,"skill-same:01M7B2K2PBDHCS5SDJHHPEV2ZW":52},{"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":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":33,"ownerHtmlUrl":34,"ownerBlogUrl":35,"ownerDescription":36,"updatedAt":37,"install":38,"resources":51},"01M7B2K2PBDHCS5SDJHHPEV2ZW","ito-inference",null,"Inspect the availability of model serving on a completed Itô compute booking and, when the canonical backend becomes available, hand off an explicitly confirmed serving manifest. Use after ito-compute has booked GPU nodes and the user asks for an OpenAI-compatible endpoint, ito-serve, hosted Kimi, or self-hosted open-weights inference. ECC implements no serving stack of its own.","lytong760-cell\u002Fcompiler_ko","lytong760-cell","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F289977458?v=4","01M7B1YD239K88PFQRD4PVC9DW","Nguyên Games","01M7B1YD2RM70TZ71QNQ0XTN4R",2,"MIT","https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fito-inference\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ito-inference\ndescription: Inspect the availability of model serving on a completed Itô compute booking and, when the canonical backend becomes available, hand off an explicitly confirmed serving manifest. Use after ito-compute has booked GPU nodes and the user asks for an OpenAI-compatible endpoint, ito-serve, hosted Kimi, or self-hosted open-weights inference. ECC implements no serving stack of its own.\nmetadata:\n  origin: ECC\n  status: scaffold\n  aliases: ito-serve, hosted-open-weights\n---\n\n# Itô Inference\n\n`ito-inference` is the sole canonical ECC skill for inference serving on Itô\ncompute. Requests naming `ito-serve` route here; do not create or install a\nsecond `ito-serve` skill. ECC never SSHes to nodes, downloads weights, launches\nan engine, or exposes an endpoint; it never books, reserves, or spends.\n\n## Current production boundary\n\nManaged serving is unavailable today. The ECC bridge exposes only `login`,\n`auth`, `find`, `status`, and explicitly gated `evals`. It has no `serve` verb.\nThe canonical runtime documents `inference` only as an unsupported compatibility\nprobe; ECC does not invoke or depend on it. The MCP surface exposes only auth,\nfind, and status. The locally enforceable guarantee is that ECC rejects `serve`\nbefore resolving or spawning the credential-bearing canonical client.\n\nTherefore stop before authentication or any command invocation. Report the\nmissing capability and return to the originating agent. Never substitute a\nlocal runner, SSH helper, browser workflow, purchase endpoint, or any untracked\nlocal `ito-serve` draft.\n\n## Required entitlement\n\nWhen serving is implemented, its first gate is a server-verified completed\nbooking. Harness memory, an RFQ, a quote, node IPs, or SSH access are not proof\nof entitlement. The backend must return fresh serving eligibility bound to the\nauthenticated account, booking, GPU topology, region, fabric, term, and model\npolicy. Expired, revoked, mismatched, incomplete, or already-released bookings\nfail closed before confirmation.\n\n## Future CLI and API contract\n\nThe intended command name is `serve`; `inference` may remain only as an\nexplicitly deprecated compatibility alias after the production contract lands.\nThe future handoff must be equivalent to:\n\n```sh\necc ito serve \\\n  --booking \u003Cserver-verified-booking-id> \\\n  --manifest \u003Cabsolute-reviewed-json-file> \\\n  --confirmation-ref \u003Copaque-non-authorizing-reference> \\\n  --idempotency-key \u003Cstable-retry-key> \\\n  --json\n```\n\nThe reviewed manifest must identify the model revision, engine and version,\nquantization, tensor\u002Fpipeline topology, endpoint exposure policy, artifact\nchecksums, storage ceiling, runtime limits, optional TTFT\u002FTPOT objectives, and\nmaximum incremental cost. No raw API key, SSH key, node password, or bearer\ntoken belongs in arguments, manifests, logs, MCP results, or chat.\n\nThe client must canonicalize the manifest path, reject symlinks, open a regular\nfile without following links, require appropriate ownership and restrictive\npermissions, enforce a bounded size, and hash bytes from the opened descriptor.\nThat digest must exactly equal the digest bound into confirmation before any\nworkload mutation. A path swap, digest mismatch, oversized file, or mutable\nunsafe file fails closed.\n\nThe canonical API—not ECC—must own workload creation and return structured JSON\nwith `ok`, `live_api_contacted`, `notice`, and either `data` or `error`. Serving\ndata must include stable booking, workload, manifest, and idempotency IDs plus a\nstate enum; it must not claim an endpoint is live until health and model checks\npass. Errors must include a stable code and safe message without secrets.\n\n## Confirmation and execution gates\n\nBefore workload creation, require all of the following:\n\n1. Fresh entitlement and serving eligibility from the canonical backend.\n2. A reviewable immutable manifest and deterministic digest.\n3. A separate single-use confirmation bound to account, action, manifest, and\n   cost, with a short expiry and replay protection. CLI arguments carry only an\n   opaque, non-authorizing confirmation reference; the server resolves and\n   consumes the bearer capability out of band.\n4. A caller-supplied idempotency key reserved atomically with the workload.\n5. Server-side fabric, capacity, model-policy, storage, and cost validation.\n\nAuthentication is identity, not workload authority. A login, API key, quote,\nor completed booking never substitutes for the serving confirmation. Inspection\nand plan generation must not create a workload. Cancel and cleanup are separate\nmutations with their own scoped confirmation and idempotency boundaries.\n\n## Lifecycle and recovery\n\nThe production surface is incomplete until the same canonical client exposes\ntenant-scoped status, logs, metrics, cancel, and cleanup operations. Every\noperation needs bounded connect and overall timeouts, revocation-aware errors,\nand structured output. After an ambiguous transport failure, query status by\nthe idempotency key before retrying; never create a second workload merely\nbecause the first response was lost. A revoked credential stops polling and\nreturns control to the originating agent without starting login automatically.\n\nOnly report `ready` after endpoint health, model identity, and canary inference\nall pass. Report intermediate and terminal failure states honestly. Cleanup must\nbe observable and must not release or modify the underlying booking unless that\nseparate economic action was explicitly authorized.\n\n## Proposed backend stages\n\nThese stages describe the future backend, not code that exists in ECC:\n\n1. Verify entitlement, topology, fabric, and cost gates.\n2. Fetch checksum-pinned weights into backend-managed storage.\n3. Emit and validate a reviewable topology\u002Fengine plan.\n4. Launch through the provider control plane, never direct root SSH from ECC.\n5. Warm up, test health and model identity, run an SLO canary, then register the\n   endpoint and redacted configuration.\n\nUntil every gate and lifecycle operation above exists in the canonical runtime,\nthis skill remains a fail-closed availability check and documentation handoff.\n","# Itô Inference\n\n`ito-inference` is the sole canonical ECC skill for inference serving on Itô\ncompute. Requests naming `ito-serve` route here; do not create or install a\nsecond `ito-serve` skill. ECC never SSHes to nodes, downloads weights, launches\nan engine, or exposes an endpoint; it never books, reserves, or spends.\n\n## Current production boundary\n\nManaged serving is unavailable today. The ECC bridge exposes only `login`,\n`auth`, `find`, `status`, and explicitly gated `evals`. It has no `serve` verb.\nThe canonical runtime documents `inference` only as an unsupported compatibility\nprobe; ECC does not invoke or depend on it. The MCP surface exposes only auth,\nfind, and status. The locally enforceable guarantee is that ECC rejects `serve`\nbefore resolving or spawning the credential-bearing canonical client.\n\nTherefore stop before authentication or any command invocation. Report the\nmissing capability and return to the originating agent. Never substitute a\nlocal runner, SSH helper, browser workflow, purchase endpoint, or any untracked\nlocal `ito-serve` draft.\n\n## Required entitlement\n\nWhen serving is implemented, its first gate is a server-verified completed\nbooking. Harness memory, an RFQ, a quote, node IPs, or SSH access are not proof\nof entitlement. The backend must return fresh serving eligibility bound to the\nauthenticated account, booking, GPU topology, region, fabric, term, and model\npolicy. Expired, revoked, mismatched, incomplete, or already-released bookings\nfail closed before confirmation.\n\n## Future CLI and API contract\n\nThe intended command name is `serve`; `inference` may remain only as an\nexplicitly deprecated compatibility alias after the production contract lands.\nThe future handoff must be equivalent to:\n\n```sh\necc ito serve \\\n  --booking \u003Cserver-verified-booking-id> \\\n  --manifest \u003Cabsolute-reviewed-json-file> \\\n  --confirmation-ref \u003Copaque-non-authorizing-reference> \\\n  --idempotency-key \u003Cstable-retry-key> \\\n  --json\n```\n\nThe reviewed manifest must identify the model revision, engine and version,\nquantization, tensor\u002Fpipeline topology, endpoint exposure policy, artifact\nchecksums, storage ceiling, runtime limits, optional TTFT\u002FTPOT objectives, and\nmaximum incremental cost. No raw API key, SSH key, node password, or bearer\ntoken belongs in arguments, manifests, logs, MCP results, or chat.\n\nThe client must canonicalize the manifest path, reject symlinks, open a regular\nfile without following links, require appropriate ownership and restrictive\npermissions, enforce a bounded size, and hash bytes from the opened descriptor.\nThat digest must exactly equal the digest bound into confirmation before any\nworkload mutation. A path swap, digest mismatch, oversized file, or mutable\nunsafe file fails closed.\n\nThe canonical API—not ECC—must own workload creation and return structured JSON\nwith `ok`, `live_api_contacted`, `notice`, and either `data` or `error`. Serving\ndata must include stable booking, workload, manifest, and idempotency IDs plus a\nstate enum; it must not claim an endpoint is live until health and model checks\npass. Errors must include a stable code and safe message without secrets.\n\n## Confirmation and execution gates\n\nBefore workload creation, require all of the following:\n\n1. Fresh entitlement and serving eligibility from the canonical backend.\n2. A reviewable immutable manifest and deterministic digest.\n3. A separate single-use confirmation bound to account, action, manifest, and\n   cost, with a short expiry and replay protection. CLI arguments carry only an\n   opaque, non-authorizing confirmation reference; the server resolves and\n   consumes the bearer capability out of band.\n4. A caller-supplied idempotency key reserved atomically with the workload.\n5. Server-side fabric, capacity, model-policy, storage, and cost validation.\n\nAuthentication is identity, not workload authority. A login, API key, quote,\nor completed booking never substitutes for the serving confirmation. Inspection\nand plan generation must not create a workload. Cancel and cleanup are separate\nmutations with their own scoped confirmation and idempotency boundaries.\n\n## Lifecycle and recovery\n\nThe production surface is incomplete until the same canonical client exposes\ntenant-scoped status, logs, metrics, cancel, and cleanup operations. Every\noperation needs bounded connect and overall timeouts, revocation-aware errors,\nand structured output. After an ambiguous transport failure, query status by\nthe idempotency key before retrying; never create a second workload merely\nbecause the first response was lost. A revoked credential stops polling and\nreturns control to the originating agent without starting login automatically.\n\nOnly report `ready` after endpoint health, model identity, and canary inference\nall pass. Report intermediate and terminal failure states honestly. Cleanup must\nbe observable and must not release or modify the underlying booking unless that\nseparate economic action was explicitly authorized.\n\n## Proposed backend stages\n\nThese stages describe the future backend, not code that exists in ECC:\n\n1. Verify entitlement, topology, fabric, and cost gates.\n2. Fetch checksum-pinned weights into backend-managed storage.\n3. Emit and validate a reviewable topology\u002Fengine plan.\n4. Launch through the provider control plane, never direct root SSH from ECC.\n5. Warm up, test health and model identity, run an SLO canary, then register the\n   endpoint and redacted configuration.\n\nUntil every gate and lifecycle operation above exists in the canonical runtime,\nthis skill remains a fail-closed availability check and documentation handoff.\n",{"name":5,"metadata":24,"description":7},{"origin":25,"status":26,"aliases":27},"ECC","scaffold","ito-serve, hosted-open-weights","main","c32dc0b68600587f78aff3310f3bbd4212173ab3",".kilo\u002Fskills\u002Fito-inference",1,"https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko","https:\u002F\u002Fko-lang.ai.studio","https:\u002F\u002Fgithub.com\u002Flytong760-cell","hanaminikata.com","hello ","2026-11-05 18:26:32",{"source":8,"skillName":5,"commands":39,"prompt":49,"zipAvailable":50},[40,43,46],{"tool":41,"command":42},"npx","npx skills add lytong760-cell\u002Fcompiler_ko --skill ito-inference",{"tool":44,"command":45},"bunx","bunx skills add lytong760-cell\u002Fcompiler_ko --skill ito-inference",{"tool":47,"command":48},"pnpm","pnpm dlx skills add lytong760-cell\u002Fcompiler_ko --skill ito-inference","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fito-inference\u002FSKILL.md",true,[],[53,60,67,74,81,88],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M7B36K22Z7155PTGEGTSDNJ4","x-api","X\u002FTwitter API integration for posting tweets, threads, reading timelines, search, and analytics. Covers OAuth auth patterns, rate limits, and platform-native content posting. Use when the user wants to interact with X programmatically.","https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fx-api\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M7B36HCAE9ESGBVS6J0FR52F","writing-for-agents","Writing documents for agents. Use when creating or editing skills, or modifying AGENTS.md or CLAUDE.md.","https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fwriting-for-agents\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M7B36DYH4ZFJ3GZHPAEMVQAY","workspace-surface-audit","Audit the active repo, MCP servers, plugins, connectors, env surfaces, and harness setup, then recommend the highest-value ECC-native skills, hooks, agents, and operator workflows. Use when the user wants help setting up Claude Code or understanding what capabilities are actually available in their environment.","https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fworkspace-surface-audit\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":19},"01M7B36A1QXMEZT8VZMFVQNR5E","wizard","Generate an interactive bash wizard that walks a human through steps only they can perform. Use when provisioning infrastructure, setting up credentials or CI secrets, walking an unfamiliar third-party dashboard, or running a one-off migration or cutover. Don't invoke this for steps the agent can perform itself.","https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fwizard\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7B364EK3N88VQQJ6HFATJQE","windows-desktop-e2e","E2E testing for Windows native desktop apps (WPF, WinForms, Win32\u002FMFC, Qt) using pywinauto and Windows UI Automation. Use when writing E2E tests for a Windows native desktop app with pywinauto or UI Automation.","https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fwindows-desktop-e2e\u002FSKILL.md",[],[],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":94,"publishedAt":19},"01M7B360YF7GGNAF21KT4QB5PG","wayfinder","Plan a huge chunk of work (more than one agent session can hold) as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.","https:\u002F\u002Fgithub.com\u002Flytong760-cell\u002Fcompiler_ko\u002Fblob\u002Fmain\u002F.kilo\u002Fskills\u002Fwayfinder\u002FSKILL.md",[],[]]