[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7809AM86V545HBFWZTKA9E6":3,"skill-same:01M7809AM86V545HBFWZTKA9E6":45},{"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":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":30,"updatedAt":31,"install":32,"resources":44},"01M7809AM86V545HBFWZTKA9E6","agent-browser-control",null,"Drive a headed Chromium via the agent-browser CLI — open\u002Fclick\u002Ftype\u002Fhover\u002Fscroll\u002Fscreenshot\u002Fsnapshot\u002Feval\u002Fpdf\u002Fnetwork-route. Use this for any web interaction; no native browser tool is exposed.","HKU-MMLab\u002FUniClawBench","HKU-MMLab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F218629317?v=4","01M7808YETMPBKR5WJNJRFSCDJ","01M7808YFS0WMXB5QAF8YG01BA",40,"MIT","https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Fdocker\u002Fbase_skills\u002Fagent-browser-control\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: agent-browser-control\ndescription: Drive a headed Chromium via the agent-browser CLI — open\u002Fclick\u002Ftype\u002Fhover\u002Fscroll\u002Fscreenshot\u002Fsnapshot\u002Feval\u002Fpdf\u002Fnetwork-route. Use this for any web interaction; no native browser tool is exposed.\nalways: true\n---\n# Agent Browser Control\n\nUse the `agent-browser` CLI for any web interaction. No other browser tool is exposed in this environment.\n\n## Binary\n\n- `agent-browser` (global, on `PATH`)\n- Chromium is pre-installed at `\u002Fusr\u002Flocal\u002Fbin\u002Fchromium`\n- A persistent browser daemon is started automatically on the first command and reused across calls\n\n## Typical workflow\n\n1. Open a page:\n\n```bash\nagent-browser open https:\u002F\u002Fexample.com\n```\n\n2. Inspect the page:\n\n```bash\nagent-browser snapshot\nagent-browser get url\nagent-browser get title\n```\n\n3. Interact with elements:\n\n```bash\nagent-browser click \"@e2\"\nagent-browser fill \"@e3\" \"text value\"\nagent-browser press Enter\nagent-browser wait --load networkidle\n```\n\nFor form submission flows, do not assume a public demo form is static just because the first click appears unchanged. After clicking the submit button:\n\n```bash\nagent-browser wait --load networkidle\nagent-browser get url\nagent-browser snapshot\n```\n\nIf needed, try the submit control again or press `Enter`, then re-check the URL and page content before concluding that submission failed.\n\nBefore clicking a final action such as a submit button, refresh the page snapshot and use the current ref from that latest snapshot. Do not rely on a stale remembered ref.\n\n4. Save evidence (screenshots go to the path you pass):\n\n```bash\nmkdir -p \u002Ftmp_workspace\u002Fresults\u002Fscreenshots\nagent-browser screenshot \u002Ftmp_workspace\u002Fresults\u002Fscreenshots\u002Fpage.png\n```\n\n`screenshot` defaults to viewport-only PNG at the exact path you provide. Pass `--full` for full-page; pass `--screenshot-format jpeg` for JPEG.\n\n5. Close the browser when done:\n\n```bash\nagent-browser close\n```\n\n## Notes\n\n- `snapshot` returns refs like `@e2`; use those refs for `click`, `fill`, and `get text`.\n- For multi-step flows, `agent-browser batch --json` can run a JSON array of commands in one invocation (daemon persists between commands anyway, so chaining with `&&` is usually enough).\n- Always wait for navigation or content changes before taking the next screenshot.\n- Save any required output files under `\u002Ftmp_workspace\u002Fresults`; the `screenshot \u003Cpath>` argument is respected literally, so point it directly at the file you want saved.\n- In benchmark tasks, write required JSON\u002Flog files incrementally as milestones are reached instead of waiting until the very end.\n","# Agent Browser Control\n\nUse the `agent-browser` CLI for any web interaction. No other browser tool is exposed in this environment.\n\n## Binary\n\n- `agent-browser` (global, on `PATH`)\n- Chromium is pre-installed at `\u002Fusr\u002Flocal\u002Fbin\u002Fchromium`\n- A persistent browser daemon is started automatically on the first command and reused across calls\n\n## Typical workflow\n\n1. Open a page:\n\n```bash\nagent-browser open https:\u002F\u002Fexample.com\n```\n\n2. Inspect the page:\n\n```bash\nagent-browser snapshot\nagent-browser get url\nagent-browser get title\n```\n\n3. Interact with elements:\n\n```bash\nagent-browser click \"@e2\"\nagent-browser fill \"@e3\" \"text value\"\nagent-browser press Enter\nagent-browser wait --load networkidle\n```\n\nFor form submission flows, do not assume a public demo form is static just because the first click appears unchanged. After clicking the submit button:\n\n```bash\nagent-browser wait --load networkidle\nagent-browser get url\nagent-browser snapshot\n```\n\nIf needed, try the submit control again or press `Enter`, then re-check the URL and page content before concluding that submission failed.\n\nBefore clicking a final action such as a submit button, refresh the page snapshot and use the current ref from that latest snapshot. Do not rely on a stale remembered ref.\n\n4. Save evidence (screenshots go to the path you pass):\n\n```bash\nmkdir -p \u002Ftmp_workspace\u002Fresults\u002Fscreenshots\nagent-browser screenshot \u002Ftmp_workspace\u002Fresults\u002Fscreenshots\u002Fpage.png\n```\n\n`screenshot` defaults to viewport-only PNG at the exact path you provide. Pass `--full` for full-page; pass `--screenshot-format jpeg` for JPEG.\n\n5. Close the browser when done:\n\n```bash\nagent-browser close\n```\n\n## Notes\n\n- `snapshot` returns refs like `@e2`; use those refs for `click`, `fill`, and `get text`.\n- For multi-step flows, `agent-browser batch --json` can run a JSON array of commands in one invocation (daemon persists between commands anyway, so chaining with `&&` is usually enough).\n- Always wait for navigation or content changes before taking the next screenshot.\n- Save any required output files under `\u002Ftmp_workspace\u002Fresults`; the `screenshot \u003Cpath>` argument is respected literally, so point it directly at the file you want saved.\n- In benchmark tasks, write required JSON\u002Flog files incrementally as milestones are reached instead of waiting until the very end.\n",{"name":5,"always":23,"description":7},true,"main","a8ac4df6c554178c339f42696e3afe688a61f4b7","docker\u002Fbase_skills\u002Fagent-browser-control",1,"https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench","https:\u002F\u002Fgithub.com\u002FHKU-MMLab","Multimedia Lab at The University of Hong Kong, a top-notch research cohort dedicated to research and educating the best minds.","2026-07-28 16:29:51",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":23},[34,37,40],{"tool":35,"command":36},"npx","npx skills add HKU-MMLab\u002FUniClawBench --skill agent-browser-control",{"tool":38,"command":39},"bunx","bunx skills add HKU-MMLab\u002FUniClawBench --skill agent-browser-control",{"tool":41,"command":42},"pnpm","pnpm dlx skills add HKU-MMLab\u002FUniClawBench --skill agent-browser-control","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Fdocker\u002Fbase_skills\u002Fagent-browser-control\u002FSKILL.md",[],[46,53,60,67,74,81],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":18},"01M78127EY3XPHMS16Y8A8B6NR","cli-anything-zoom","Command-line interface for Zoom — manage meetings, participants, and recordings from the command line via the Zoom REST API.","https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Finjection\u002F205_cross_platform_zh\u002Ftask_205_33_calendar_scheduling\u002Fskills\u002Fcli-anything-zoom\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":18},"01M78123F7PCSTEZ8CT3DBEGBT","netease-music-cli","CLI for NetEase Cloud Music — search songs\u002Falbums\u002Fplaylists, get lyrics, and manage your music library from the terminal.","https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Finjection\u002F205_cross_platform_zh\u002Ftask_205_26_music_crossid\u002Fskills\u002Fnetease-music-cli\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":18},"01M78120M9QEN0MQB9X0FCKSZX","google-calendar-gcalcli","Install and use gcalcli for live Google Calendar read access using the provided OAuth cache.","https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Finjection\u002F203_long_context_zh\u002Ftask_203_08_hk_jpop_calendar_ticketing\u002Fskills\u002Fgoogle-calendar-gcalcli\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":18},"01M7811TTCHD4JW4FG5HG1PVY9","Excel \u002F XLSX","Create, inspect, and edit Microsoft Excel workbooks and XLSX files with reliable formulas, dates, types, formatting, recalculation, and template preservation. Use when (1) the task is about Excel, `.xlsx`, `.xlsm`, `.xls`, `.csv`, or `.tsv`; (2) formulas, formatting, workbook structure, or compatibility matter; (3) the file must stay reliable after edits.","https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Finjection\u002F201_skill_usage_zh\u002Ftask_201_40_excel_dashboard_qa\u002Fskills\u002Fexcel-xlsx\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":18},"01M7811PGJQ4XDQ7XJ0TKNSEE6","duckdb-en","DuckDB CLI specialist for SQL analysis, data processing and file conversion. Use for SQL queries, CSV\u002FParquet\u002FJSON analysis, database queries, or data conversion. Triggers on \"duckdb\", \"sql\", \"query\", \"data analysis\", \"parquet\", \"convert data\".","https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Finjection\u002F201_skill_usage_zh\u002Ftask_201_40_excel_dashboard_qa\u002Fskills\u002Fduckdb-cli-ai-skills\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":18},"01M7811JSJW78171B25VJ285A0","automate-excel","Automates reading, writing, merging, transforming, and validating Excel (.xlsx\u002F.xls) files. Use when the user works with spreadsheets, .xlsx files, Excel data — for cleaning, joining, validating, deduplicating, aggregating, filtering, or auditing tabular data.","https:\u002F\u002Fgithub.com\u002FHKU-MMLab\u002FUniClawBench\u002Fblob\u002Fmain\u002Finjection\u002F201_skill_usage_zh\u002Ftask_201_40_excel_dashboard_qa\u002Fskills\u002Fautomate-excel\u002FSKILL.md",[],[]]