[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6CPDVWF84E2NEPFCTES8D8R":3,"skill-same:01M6CPDVWF84E2NEPFCTES8D8R":115},{"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":28,"publishedAt":49,"linkOnly":50,"markdownContent":51,"markdownBody":52,"frontMatter":53,"refName":68,"commitSha":69,"skillDirectory":70,"forksCount":71,"repositoryHtmlUrl":72,"repositoryHomepageUrl":73,"ownerHtmlUrl":74,"ownerBlogUrl":75,"ownerDescription":76,"updatedAt":77,"install":78,"resources":90},"01M6CPDVWF84E2NEPFCTES8D8R","browser_session",null,"Multi-step browser automation via browser-use — the open-source\nframework for autonomous browser navigation (91K+ stars, MIT). Use\nINSTEAD of `web` when the task spans more than a single page fetch:\n\n  - Login + navigate + extract (auth state preserved across steps)\n  - 5+ page interaction sequences\n  - DOM-numbered element targeting (the framework shows the agent\n    which element index to click — far more reliable than guessing\n    CSS selectors)\n  - JS-heavy SPAs that need real interaction state, not just\n    rendered HTML\n\nCost: cold start ~10-20s (browser warmup + LLM init); per-step\n~2-5s. **Don't use for one-shot fetches** — `web` is faster and\ncheaper for those. Use this when the task description naturally\ncontains \"log in\", \"navigate to\", \"fill the form\", \"find on page X\nthen click\", or any sentence with multiple verbs of interaction.\n\nFirst-time setup (one machine, ~5 min):\n  cd skills\u002Fbrowser_session\n  .\u002Fsetup.sh\n\nThis creates a per-skill venv at .\u002F.venv\u002F and installs browser-use\n+ Chromium. Subsequent calls use the venv directly — zero pollution\nof the system Python.\n\nLLM selection (env-driven, in order of preference):\n  ANTHROPIC_API_KEY         → Claude\n  OPENAI_API_KEY            → GPT-4o\n  OLLAMA_BASE_URL=URL       → local Ollama via OpenAI-compat\nOptional: BROWSER_USE_MODEL=\u003Cname> to override the model.\nOptional: BROWSER_USE_HEADLESS=false to watch the browser run.","LocalKinAI\u002Fkinclaw","LocalKinAI","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F228881863?v=4","01M6CPDM368JVRW78EPV2NDABD","01M6CPDM7CKTKCXE8JD7K3658Q",9,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fbrowser_session\u002FSKILL.md",[17,22,25],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000009","前端开发","SCENARIO","使用场景",{"categoryNo":23,"categoryName":24,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":26,"categoryName":27,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发",[29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48],"browser session","browser","session","model","runner","selection","log","autonomous","fetch","browser navigation","ask","css","plan","open","api","SET","search","init","auth","browser automation","2026-11-10 10:12:30",false,"---\nname: browser_session\ndescription: |\n  Multi-step browser automation via browser-use — the open-source\n  framework for autonomous browser navigation (91K+ stars, MIT). Use\n  INSTEAD of `web` when the task spans more than a single page fetch:\n\n    - Login + navigate + extract (auth state preserved across steps)\n    - 5+ page interaction sequences\n    - DOM-numbered element targeting (the framework shows the agent\n      which element index to click — far more reliable than guessing\n      CSS selectors)\n    - JS-heavy SPAs that need real interaction state, not just\n      rendered HTML\n\n  Cost: cold start ~10-20s (browser warmup + LLM init); per-step\n  ~2-5s. **Don't use for one-shot fetches** — `web` is faster and\n  cheaper for those. Use this when the task description naturally\n  contains \"log in\", \"navigate to\", \"fill the form\", \"find on page X\n  then click\", or any sentence with multiple verbs of interaction.\n\n  First-time setup (one machine, ~5 min):\n    cd skills\u002Fbrowser_session\n    .\u002Fsetup.sh\n\n  This creates a per-skill venv at .\u002F.venv\u002F and installs browser-use\n  + Chromium. Subsequent calls use the venv directly — zero pollution\n  of the system Python.\n\n  LLM selection (env-driven, in order of preference):\n    ANTHROPIC_API_KEY         → Claude\n    OPENAI_API_KEY            → GPT-4o\n    OLLAMA_BASE_URL=URL       → local Ollama via OpenAI-compat\n  Optional: BROWSER_USE_MODEL=\u003Cname> to override the model.\n  Optional: BROWSER_USE_HEADLESS=false to watch the browser run.\ncommand:\n  - sh\n  - -c\n  - |\n    TASK=\"$1\"\n    [ -z \"$TASK\" ] && { echo \"task required\" >&2; exit 1; }\n\n    # cwd is set to the skill dir by the kernel.\n    if [ ! -x .\u002F.venv\u002Fbin\u002Fpython ]; then\n      echo \"browser_session: venv missing — first-time setup required:\" >&2\n      echo \"  cd skills\u002Fbrowser_session && .\u002Fsetup.sh\" >&2\n      exit 1\n    fi\n\n    if ! .\u002F.venv\u002Fbin\u002Fpython -c \"import browser_use\" 2>\u002Fdev\u002Fnull; then\n      echo \"browser_session: venv exists but browser-use not installed —\" >&2\n      echo \"  re-run: cd skills\u002Fbrowser_session && .\u002Fsetup.sh\" >&2\n      exit 1\n    fi\n\n    # Sensible defaults so the skill works out of the box on a typical\n    # KinClaw setup (local Ollama running, no external API keys\n    # required). User-set env (ANTHROPIC_API_KEY etc.) still takes\n    # precedence — runner.py picks Anthropic > OpenAI > Ollama.\n    : \"${OLLAMA_BASE_URL:=http:\u002F\u002Flocalhost:11434}\"\n    : \"${BROWSER_USE_MODEL:=kimi-k2.6:cloud}\"\n    export OLLAMA_BASE_URL BROWSER_USE_MODEL\n\n    .\u002F.venv\u002Fbin\u002Fpython .\u002Frunner.py \"$TASK\"\n  - \"_\"\nargs:\n  - \"{{task}}\"\nschema:\n  task:\n    type: string\n    description: |\n      High-level natural language task description. browser-use's\n      internal LLM agent will plan + execute the steps. Examples:\n\n      \"Open Hacker News, find the top story right now, return its\n       title + URL + first paragraph of the article\"\n\n      \"Search GitHub for repositories matching 'computer use agent',\n       sorted by stars, return the top 5 with their star counts\"\n\n      \"Go to weather.com for 94025, extract today's high\u002Flow and the\n       7-day forecast as a markdown table\"\n\n      The task should be self-contained — browser-use can't ask\n      followup questions, it just plans and acts.\n    required: true\ntimeout: 600\n---\n\n# browser_session — multi-step browser automation\n\nA `web` claw on steroids. While `web` is a one-shot Playwright\nwrapper good for \"fetch this page and read it\", `browser_session`\nhands the entire interaction loop to [browser-use](https:\u002F\u002Fgithub.com\u002Fbrowser-use\u002Fbrowser-use)\n— a framework that keeps a persistent browser, decorates the DOM\nwith numbered element IDs, and has its own LLM-driven planning loop\nto execute multi-step tasks.\n\n## When to use which\n\n| Need | Tool | Why |\n|---|---|---|\n| Read one page's content | `web` | Cheap (~3s cold start), trivial |\n| Run one JS expression | `web` | Same |\n| Take one screenshot | `web` | Same |\n| Click + type once | `web` | Same — `click=` + `type_text=` |\n| Login + then do stuff | `browser_session` | Persistent session |\n| 5+ page navigation | `browser_session` | No cold-start per step |\n| Form + multi-page wizard | `browser_session` | Framework handles state |\n| Anything where a CSS selector won't survive | `browser_session` | Element-index targeting |\n\n## Why a super-skill\n\n`browser_session` is the first member of LocalKin's **super-skill**\npattern: skills that wrap a battle-tested third-party OSS framework\n(here, browser-use's 91K-star codebase) inside a thin SKILL.md\nadapter. The kinclaw kernel doesn't care that there's an entire\nLLM-driven planning agent inside; it just sees one tool call going\nin, one result coming out.\n\nThis is \"thin soul, fat skill\" pushed to its useful extreme:\n\n- We don't reinvent browser-use. We host it.\n- Other LocalKin souls (researcher \u002F marketer \u002F curator) can also\n  enable `browser_session` and get the same capability.\n- If browser-use ships a new version, we re-run `setup.sh` — the\n  SKILL.md interface stays the same.\n\n## Setup details\n\n`.\u002Fsetup.sh`:\n\n1. Picks a Python ≥3.11 (homebrew, pyenv, or system) — browser-use\n   doesn't support 3.10\n2. Creates `.\u002F.venv\u002F` (per-skill venv, isolated)\n3. `pip install browser-use playwright`\n4. `playwright install chromium` (~92 MB Chrome Headless Shell)\n5. Verifies the imports work\n\nThe venv lives in `.\u002F.venv\u002F`. **Add `.venv\u002F` to .gitignore** so the\n~500 MB of dependencies don't get committed.\n\n## LLM selection\n\nThe framework drives a real LLM internally. Pick one with vision\nsupport — browser-use feeds screenshots to the LLM for visual\nreasoning. Recommended (in descending preference):\n\n1. **`claude-sonnet-4.5`** via `ANTHROPIC_API_KEY` — current SOTA\n   for browser-use's WebVoyager benchmark\n2. **`gpt-4o`** via `OPENAI_API_KEY` — second-best, cheaper\n3. **Local Ollama** with a vision model (e.g. `qwen2.5-vl:7b`) via\n   `OLLAMA_BASE_URL=http:\u002F\u002Flocalhost:11434` — free but ~3-5x slower\n   per step; good for long batch runs you don't want to pay for\n\nSet `BROWSER_USE_MODEL=\u003Cname>` to override the default for the\nselected provider.\n\n## Returns\n\nstdout: the agent's final answer (a string — could be markdown if\nthe task asked for structured output).\n\nstderr: errors, warnings, the playwright\u002Fchromium boot noise.\n\nNon-zero exit: agent failed or reported errors. The kernel will\nsurface this as a tool error to the calling soul.\n\n## Costs\n\nA typical 5-step task with Claude Sonnet:\n- ~30-60 seconds wall-clock\n- ~$0.05-0.15 in API tokens (browser-use sends DOM + screenshot at\n  every step)\n\nFor dev\u002Ftest use the local Ollama path; for production runs use\nAnthropic.\n","# browser_session — multi-step browser automation\n\nA `web` claw on steroids. While `web` is a one-shot Playwright\nwrapper good for \"fetch this page and read it\", `browser_session`\nhands the entire interaction loop to [browser-use](https:\u002F\u002Fgithub.com\u002Fbrowser-use\u002Fbrowser-use)\n— a framework that keeps a persistent browser, decorates the DOM\nwith numbered element IDs, and has its own LLM-driven planning loop\nto execute multi-step tasks.\n\n## When to use which\n\n| Need | Tool | Why |\n|---|---|---|\n| Read one page's content | `web` | Cheap (~3s cold start), trivial |\n| Run one JS expression | `web` | Same |\n| Take one screenshot | `web` | Same |\n| Click + type once | `web` | Same — `click=` + `type_text=` |\n| Login + then do stuff | `browser_session` | Persistent session |\n| 5+ page navigation | `browser_session` | No cold-start per step |\n| Form + multi-page wizard | `browser_session` | Framework handles state |\n| Anything where a CSS selector won't survive | `browser_session` | Element-index targeting |\n\n## Why a super-skill\n\n`browser_session` is the first member of LocalKin's **super-skill**\npattern: skills that wrap a battle-tested third-party OSS framework\n(here, browser-use's 91K-star codebase) inside a thin SKILL.md\nadapter. The kinclaw kernel doesn't care that there's an entire\nLLM-driven planning agent inside; it just sees one tool call going\nin, one result coming out.\n\nThis is \"thin soul, fat skill\" pushed to its useful extreme:\n\n- We don't reinvent browser-use. We host it.\n- Other LocalKin souls (researcher \u002F marketer \u002F curator) can also\n  enable `browser_session` and get the same capability.\n- If browser-use ships a new version, we re-run `setup.sh` — the\n  SKILL.md interface stays the same.\n\n## Setup details\n\n`.\u002Fsetup.sh`:\n\n1. Picks a Python ≥3.11 (homebrew, pyenv, or system) — browser-use\n   doesn't support 3.10\n2. Creates `.\u002F.venv\u002F` (per-skill venv, isolated)\n3. `pip install browser-use playwright`\n4. `playwright install chromium` (~92 MB Chrome Headless Shell)\n5. Verifies the imports work\n\nThe venv lives in `.\u002F.venv\u002F`. **Add `.venv\u002F` to .gitignore** so the\n~500 MB of dependencies don't get committed.\n\n## LLM selection\n\nThe framework drives a real LLM internally. Pick one with vision\nsupport — browser-use feeds screenshots to the LLM for visual\nreasoning. Recommended (in descending preference):\n\n1. **`claude-sonnet-4.5`** via `ANTHROPIC_API_KEY` — current SOTA\n   for browser-use's WebVoyager benchmark\n2. **`gpt-4o`** via `OPENAI_API_KEY` — second-best, cheaper\n3. **Local Ollama** with a vision model (e.g. `qwen2.5-vl:7b`) via\n   `OLLAMA_BASE_URL=http:\u002F\u002Flocalhost:11434` — free but ~3-5x slower\n   per step; good for long batch runs you don't want to pay for\n\nSet `BROWSER_USE_MODEL=\u003Cname>` to override the default for the\nselected provider.\n\n## Returns\n\nstdout: the agent's final answer (a string — could be markdown if\nthe task asked for structured output).\n\nstderr: errors, warnings, the playwright\u002Fchromium boot noise.\n\nNon-zero exit: agent failed or reported errors. The kernel will\nsurface this as a tool error to the calling soul.\n\n## Costs\n\nA typical 5-step task with Claude Sonnet:\n- ~30-60 seconds wall-clock\n- ~$0.05-0.15 in API tokens (browser-use sends DOM + screenshot at\n  every step)\n\nFor dev\u002Ftest use the local Ollama path; for production runs use\nAnthropic.\n",{"args":54,"name":5,"schema":56,"command":61,"timeout":66,"description":67},[55],"{{task}}",{"task":57},{"type":58,"required":59,"description":60},"string",true,"High-level natural language task description. browser-use's\ninternal LLM agent will plan + execute the steps. Examples:\n\n\"Open Hacker News, find the top story right now, return its\n title + URL + first paragraph of the article\"\n\n\"Search GitHub for repositories matching 'computer use agent',\n sorted by stars, return the top 5 with their star counts\"\n\n\"Go to weather.com for 94025, extract today's high\u002Flow and the\n 7-day forecast as a markdown table\"\n\nThe task should be self-contained — browser-use can't ask\nfollowup questions, it just plans and acts.\n",[62,63,64,65],"sh","-c","TASK=\"$1\"\n[ -z \"$TASK\" ] && { echo \"task required\" >&2; exit 1; }\n\n# cwd is set to the skill dir by the kernel.\nif [ ! -x .\u002F.venv\u002Fbin\u002Fpython ]; then\n  echo \"browser_session: venv missing — first-time setup required:\" >&2\n  echo \"  cd skills\u002Fbrowser_session && .\u002Fsetup.sh\" >&2\n  exit 1\nfi\n\nif ! .\u002F.venv\u002Fbin\u002Fpython -c \"import browser_use\" 2>\u002Fdev\u002Fnull; then\n  echo \"browser_session: venv exists but browser-use not installed —\" >&2\n  echo \"  re-run: cd skills\u002Fbrowser_session && .\u002Fsetup.sh\" >&2\n  exit 1\nfi\n\n# Sensible defaults so the skill works out of the box on a typical\n# KinClaw setup (local Ollama running, no external API keys\n# required). User-set env (ANTHROPIC_API_KEY etc.) still takes\n# precedence — runner.py picks Anthropic > OpenAI > Ollama.\n: \"${OLLAMA_BASE_URL:=http:\u002F\u002Flocalhost:11434}\"\n: \"${BROWSER_USE_MODEL:=kimi-k2.6:cloud}\"\nexport OLLAMA_BASE_URL BROWSER_USE_MODEL\n\n.\u002F.venv\u002Fbin\u002Fpython .\u002Frunner.py \"$TASK\"\n","_",600,"Multi-step browser automation via browser-use — the open-source\nframework for autonomous browser navigation (91K+ stars, MIT). Use\nINSTEAD of `web` when the task spans more than a single page fetch:\n\n  - Login + navigate + extract (auth state preserved across steps)\n  - 5+ page interaction sequences\n  - DOM-numbered element targeting (the framework shows the agent\n    which element index to click — far more reliable than guessing\n    CSS selectors)\n  - JS-heavy SPAs that need real interaction state, not just\n    rendered HTML\n\nCost: cold start ~10-20s (browser warmup + LLM init); per-step\n~2-5s. **Don't use for one-shot fetches** — `web` is faster and\ncheaper for those. Use this when the task description naturally\ncontains \"log in\", \"navigate to\", \"fill the form\", \"find on page X\nthen click\", or any sentence with multiple verbs of interaction.\n\nFirst-time setup (one machine, ~5 min):\n  cd skills\u002Fbrowser_session\n  .\u002Fsetup.sh\n\nThis creates a per-skill venv at .\u002F.venv\u002F and installs browser-use\n+ Chromium. Subsequent calls use the venv directly — zero pollution\nof the system Python.\n\nLLM selection (env-driven, in order of preference):\n  ANTHROPIC_API_KEY         → Claude\n  OPENAI_API_KEY            → GPT-4o\n  OLLAMA_BASE_URL=URL       → local Ollama via OpenAI-compat\nOptional: BROWSER_USE_MODEL=\u003Cname> to override the model.\nOptional: BROWSER_USE_HEADLESS=false to watch the browser run.\n","main","d1b92d10cddb779854af342bea74d0b6c46788e2","skills\u002Fbrowser_session",3,"https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw","https:\u002F\u002Fwww.localkin.app\u002Fkinclaw","https:\u002F\u002Fgithub.com\u002FLocalKinAI","localkin.ai","Faith and Chinese medicine at localkin.ai · the LocalKin agent platform at localkin.dev · open-source tools (kinfer, KinClaw, …) at localkin.app","2026-11-01 12:25:06",{"source":8,"skillName":5,"commands":79,"prompt":89,"zipAvailable":59},[80,83,86],{"tool":81,"command":82},"npx","npx skills add LocalKinAI\u002Fkinclaw --skill browser_session",{"tool":84,"command":85},"bunx","bunx skills add LocalKinAI\u002Fkinclaw --skill browser_session",{"tool":87,"command":88},"pnpm","pnpm dlx skills add LocalKinAI\u002Fkinclaw --skill browser_session","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fbrowser_session\u002FSKILL.md",[91,97,103,107,111],{"path":92,"resourceType":93,"resourceTypeDesc":94,"sizeBytes":95,"sourceHtmlUrl":96},"references\u002Fcore\u002Fguide.md","REFERENCE","参考文档",2510,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FLVTD-LLC\u002Fskills\u002Fskills\u002Fgame-balance-economy\u002Freferences\u002Fcore\u002Fguide.md",{"path":98,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":101,"sourceHtmlUrl":102},"runner.py","SCRIPT","脚本",6238,"https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fbrowser_session\u002Frunner.py",{"path":104,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":105,"sourceHtmlUrl":106},"scripts\u002Fexpected_value.py",3402,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FLVTD-LLC\u002Fskills\u002Fskills\u002Fgame-balance-economy\u002Fscripts\u002Fexpected_value.py",{"path":108,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":109,"sourceHtmlUrl":110},"setup.sh",2196,"https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fbrowser_session\u002Fsetup.sh",{"path":112,"resourceType":93,"resourceTypeDesc":94,"sizeBytes":113,"sourceHtmlUrl":114},"workflows\u002Ftuning-pass.md",887,"https:\u002F\u002Fgithub.com\u002Fhashgraph-online\u002Fawesome-codex-plugins\u002Fblob\u002Fmain\u002Fplugins\u002FLVTD-LLC\u002Fskills\u002Fskills\u002Fgame-balance-economy\u002Fworkflows\u002Ftuning-pass.md",[116,146,160,187,198,209],{"skillNo":117,"name":118,"nameZh":6,"description":119,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":120,"categories":121,"keywords":129,"publishedAt":49},"01M6CPGNC9S31FTGMV90NZ0EN9","web","Universal web tool — fetches and interacts with web pages through\nreal Chromium (Playwright). Renders JavaScript, follows redirects,\nhandles cookies, bypasses many anti-scrape stubs that block plain\n`web_fetch`. Each call launches a fresh browser, executes the\nrequested flow, closes. ~2-3s cold start; no persistent state.\n\nUse INSTEAD of `web_fetch` when:\n  - Target is a JS-heavy SPA (prices via XHR, lazy content)\n  - `web_fetch` returned an anti-scrape placeholder (Zhihu's\n    \"click here\" stub, JD's chrome-only HTML)\n  - You need to interact (click, fill input) before extracting\n\nCommon patterns:\n  fetch rendered text:           url=X\n  wait for content then read:    url=X wait_for=\".price\"\n  extract specific element:      url=X selector=\".product-list\"\n  click and read result:         url=X click=\".search-button\"\n  fill form, then read:          url=X click=\"input[name=q]\" type_text=\"kinclaw\"\n  fill + Enter (React forms):    url=X click=\"input\" type_text=\"hi\" press_enter=true wait_for=\".reply\"\n  screenshot the page:           url=X screenshot=true (returns image:\u002F\u002F marker)\n  run JS and get result:         url=X js=\"document.title\"\n\nFirst-time setup (one-time per machine, ~500MB Chromium download):\n  pip install playwright\n  playwright install chromium","https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fweb\u002FSKILL.md",[122,125,128],{"categoryNo":123,"categoryName":124,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000002","文档处理",{"categoryNo":126,"categoryName":127,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},[118,130,45,131,132,35,133,134,135,136,137,138,139,140,141,142,143,144,40,145],"image","forms","files","read","text","wait for","how","scrape","start","PDF","run","generate","placeholder","change","react","directory",{"skillNo":147,"name":148,"nameZh":6,"description":149,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":150,"categories":151,"keywords":152,"publishedAt":49},"01M6CPGJTGF7P8QS3HXPCZE3KV","weather","Current conditions plus a three-day forecast (today \u002F tomorrow \u002F day\nafter) for a place, in metric, from wttr.in. No API key. Use this for\nany weather question — including \"明天\" and \"后天\" — before reaching\nfor web search.","https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fweather\u002FSKILL.md",[],[148,45,153,154,155,156,157,158,43,159,118],"three","location","use","question","python3","web search","forecast",{"skillNo":161,"name":162,"nameZh":6,"description":163,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":164,"categories":165,"keywords":170,"publishedAt":49},"01M6CPGDR0W8FR0WMBBJ2282GM","tts","Synthesize speech from text and play it through the macOS speakers.\nTalks to a local LocalKin Service Audio Server (default :8001 \u002F Kokoro).\nWhen `record` is running with audio=true, the spoken audio is\ncaptured into the recording — use this in place of `shell say` for\nhigh-quality multilingual narration in demo videos. Set the\nTTS_ENDPOINT env var to point at a different server.","https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Ftts\u002FSKILL.md",[166,169],{"categoryNo":167,"categoryName":168,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000016","图像与多媒体",{"categoryNo":26,"categoryName":27,"taxonomyType":20,"taxonomyTypeDesc":21},[162,171,172,173,174,175,176,177,178,179,180,32,134,181,182,44,183,184,185,186],"next","back","chinese","multilingual","default","Echo","payload","pick","code","auto","Audio","Video","data","server","only","macos",{"skillNo":188,"name":189,"nameZh":6,"description":190,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":191,"categories":192,"keywords":193,"publishedAt":49},"01M6CPGAM9K2VN0MVYD4CMC92E","translate","Translate text between languages using the system's translate command","https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Ftranslate\u002FSKILL.md",[],[189,173,194,195,176,196,134,197],"spanish","language","command","system",{"skillNo":199,"name":200,"nameZh":6,"description":201,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":202,"categories":203,"keywords":207,"publishedAt":49},"01M6CPG8DWEY0GWCFDGTTBWA2Q","summarize","Summarize a file's contents into key bullet points","https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fsummarize\u002FSKILL.md",[204],{"categoryNo":205,"categoryName":206,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000021","资料检索与研究",[200,208],"file",{"skillNo":210,"name":211,"nameZh":6,"description":212,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":213,"categories":214,"keywords":216,"publishedAt":49},"01M6CPG48P3AYKRREAGYSGT944","stt","Transcribe an audio file (wav\u002Fmp3\u002Fmp4\u002Fm4a) into text via a local\nLocalKin Service Audio Server (default :8000 \u002F SenseVoice). Pair\nwith `record action=start mic=true` to turn a recorded mic track\ninto text. Set STT_ENDPOINT env var to point at a different server.","https:\u002F\u002Fgithub.com\u002FLocalKinAI\u002Fkinclaw\u002Fblob\u002Fmain\u002Fskills\u002Fstt\u002FSKILL.md",[215],{"categoryNo":167,"categoryName":168,"taxonomyType":20,"taxonomyTypeDesc":21},[211,134,172,217,175,176,218,180,219,220,221,222,181,223,44,138,184,224,179,225],"do","en","pair","just","returns","mp4","endpoint","transcribe","go"]