[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M79JX9KW0Z2HWEAVKC2Z693J":3,"skill-same:01M79JX9KW0Z2HWEAVKC2Z693J":43},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"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":23,"commitSha":24,"skillDirectory":25,"forksCount":13,"repositoryHtmlUrl":26,"repositoryHomepageUrl":6,"ownerHtmlUrl":27,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":28,"install":29,"resources":42},"01M79JX9KW0Z2HWEAVKC2Z693J","browser-automation",null,"Use when controlling web pages with the OpenClaw browser tool, especially multi-step flows, login checks, tab management, or recovery from stale refs\u002Ftimeouts.","leon36000\u002FOPSYS-AI","leon36000","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F216249790?v=4","01M79JWJK00J10J79XZA3G4KH0","01M79JWJKTT8J3C7PZ0TX7X4PQ",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Fcore\u002Fextensions\u002Fbrowser\u002Fskills\u002Fbrowser-automation\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: browser-automation\ndescription: Use when controlling web pages with the OpenClaw browser tool, especially multi-step flows, login checks, tab management, or recovery from stale refs\u002Ftimeouts.\nuser-invocable: false\n---\n\n# Browser Automation\n\nUse this skill when you need the `browser` tool for anything beyond a single page check.\n\n## Operating Loop\n\n1. Check browser state before acting:\n   - `openclaw browser doctor` or `action=\"status\"` when the browser\u002Fplugin setup itself may be broken.\n   - `action=\"status\"` for availability.\n   - `action=\"profiles\"` if login state or profile choice matters.\n   - `action=\"tabs\"` before opening a new tab if retries\u002Ftimeouts may have left windows behind.\n2. Prefer stable tab handles:\n   - Open important tabs with `label`, for example `label=\"meet\"`.\n   - After `action=\"tabs\"` or `action=\"open\"`, store `suggestedTargetId` and pass it as `targetId` in later calls.\n   - `suggestedTargetId` is the label when one exists, otherwise the stable `tabId` handle like `t1`.\n   - Avoid relying on raw DevTools `targetId` except for immediate diagnostics; it can change under Chromium target replacement.\n3. Read before you click:\n   - For “read the page and answer X,” use `action=\"text\"` with optional `selector` and `maxChars` for bounded visible prose (first selector match, otherwise article\u002Fmain\u002Fbody). On existing-session profiles, use `snapshot` instead. Efficient snapshots omit most prose.\n   - For virtualized lists, scroll through each segment, capture only the relevant rows, then merge the results.\n   - Use `action=\"snapshot\"` on the intended `targetId`.\n   - Add snapshot `query` to find lines containing all query tokens, ignoring case; matching lines keep their refs.\n   - Use the same `targetId` for follow-up actions so refs stay on the same tab.\n   - For durable Playwright refs, request `refs=\"aria\"` when supported. If you receive `axN` refs from `snapshotFormat=\"aria\"`, use them only after that same snapshot call; stale or unbound `axN` refs fail fast and need a fresh snapshot.\n   - Use `urls=true` when link text is ambiguous or a direct navigation target would avoid brittle clicks.\n   - Use `labels=true` on snapshot or screenshot when visual position matters. On Playwright-backed profiles, the response includes an `annotations` array (`{ref, number, role, name?, box}`) with each ref's bounding box in the captured image's coordinate space, so you can reason about position without re-snapshotting; screenshot labels can also combine with `fullPage=true` (CLI: `--full-page`) to label the whole document, or `ref` \u002F `element` to clip to one element. `profile=\"user\"` and other existing-session (chrome-mcp) profiles render an overlay into page screenshots but do not attach `annotations` or use the Playwright full-page\u002Fref\u002Felement projection helper, so read positions from the labeled image itself on those profiles. The raw-CDP fallback (no Playwright) does not support labeled screenshots at all and returns a 501, so only request `labels` when Playwright is available.\n4. Act narrowly:\n   - Prefer `action=\"act\"` with a ref from the latest snapshot.\n   - `navigate` returns the loaded page's compact snapshot inline, and batch `act` results that report a cross-document navigation include fresh page state; use those refs directly instead of a follow-up snapshot call.\n   - After a single act that triggers navigation, and after modal changes or form submissions, snapshot again before the next action.\n   - Avoid blind waits. Wait for visible UI state when possible.\n   - Use `action=\"emulate\"` with `device`, `colorScheme`, `timezoneId`, or `locale` when testing those settings; snapshot again afterward. Existing-session profiles do not support emulation.\n5. Report real blockers:\n   - Debug network failures with `action=\"requests\"`, optional URL\u002Ftype `filter`, and `limit` (default 50 recent entries). `clear=true` clears the collected log after reading. Use a managed profile; existing-session profiles do not support this log.\n   - Debug page errors with `action=\"errors\"` and `limit` (default 50 recent entries). `clear=true` clears the collected log after reading. Existing-session profiles do not support this log.\n   - If the page needs login, permission, captcha, 2FA, camera\u002Fmicrophone approval, or another manual step, stop and tell the user exactly what is needed.\n   - Do not claim the browser is not logged in just because the current page shows a permission or onboarding dialog. Inspect the visible UI first.\n\n## Browser batch CLI\n\n`openclaw browser batch` runs an array of nested `\u002Fact` actions in one `\u002Fact` call (the same `kind=\"batch\"` runtime reached through the agent tool), so CLI users and scripts can combine actions like `wait`, `click`, `type`, and `evaluate` into a single replayable plan without per-action round trips. Each entry in `actions[]` is a `BrowserActRequest` — the closed union the `\u002Fact` route accepts — not arbitrary `openclaw browser` subcommands. `batch` is not supported on `profile=\"user\"` and other existing-session (chrome-mcp) profiles; send actions individually there.\n\n- CLI: `openclaw browser batch --actions '\u003Cjson>'`, `--actions-file plan.json`, or `--actions-file -` for stdin. `--actions-file` and stdin input are capped at 1,000,000 bytes; split larger plans into multiple batch commands. `--continue` sets `stopOnError=false`; default stops on first error.\n- Ref lifecycle: refs come from a `snapshot` run before the batch (snapshot is not a nested action). A nested action that changes page state — such as a `click` that triggers navigation, or an `evaluate` that mutates the DOM — can invalidate earlier refs for the rest of the batch; put state-changing actions first, or split into a follow-up batch after re-snapshotting. Navigation and re-snapshotting happen outside the batch, since `open`, `navigate`, and `snapshot` are not `\u002Fact` kinds.\n- Target id: nested actions share the request's tab; an explicit nested `targetId` that resolves to a different tab is rejected with `ACT_TARGET_ID_MISMATCH`.\n- Response: `{ \"results\": [{ \"ok\": true } | { \"ok\": false, \"error\": \"...\" }, ...] }` in order; with default `stopOnError` the array ends at the first failure. Any failed entry exits nonzero; use `--json` to preserve the full response in scripts.\n\n## Code Mode Loop\n\nWhen `tools.codeMode` is enabled, the Browser tool has no normal turn — it is cataloged behind `exec`\u002F`wait`. Call it from exec cells as an async global, using the callable name the exec quick index advertises for the Browser tool (normally `browser`; colliding names get suffixed, and a client tool can win an identical name). An exact `catalog.search(\"browser\")` returns a handle already bound to the effective callable name, so resolve the handle in each cell and call it instead of hard-coding the literal global; an empty result means the Browser tool is not cataloged in this run.\n\nKeep the same labeled tab through the loop, and alternate reads with actions. Each `exec` cell starts a fresh VM — bindings from a completed cell are gone in the next, and only runs left `waiting` keep their state until `wait` resumes them — so carry comparison state across cells by returning it and re-embedding the returned values in the next cell:\n\n```javascript\n\u002F\u002F previous = the url\u002FnewElements returned by the last completed cell (a fresh\n\u002F\u002F VM runs this cell, so prior bindings do not exist here).\nconst previous = { url: \"https:\u002F\u002Fexample.com\u002Finbox\", newElements: 0 };\nconst [browser] = await catalog.search(\"browser\", { limit: 1 });\nconst details = await browser({\n  action: \"snapshot\",\n  snapshotFormat: \"ai\",\n  targetId: \"task\",\n  refs: \"aria\",\n  interactive: true,\n});\nconst changed =\n  details?.url !== previous.url ||\n  (details?.newElements ?? 0) > 0 ||\n  details?.blockedByDialog === true;\nreturn {\n  targetId: details?.targetId,\n  url: details?.url,\n  newElements: details?.newElements,\n  stats: details?.stats,\n  changed,\n};\n```\n\n- Code-mode calls return the tool's structured `details` directly (`targetId`, `url`, `newElements`, `stats`, `blockedByDialog`); rendered page text is not returned to code cells.\n- To read text inside code mode, run a targeted `act` evaluate (requires the evaluate capability; `browser.evaluateEnabled` can disable it) and keep the returned value bounded, because page-script output is untrusted:\n\n```javascript\nconst [browser] = await catalog.search(\"browser\", { limit: 1 });\nconst read = await browser({\n  action: \"act\",\n  kind: \"evaluate\",\n  fn: \"() => document.body.innerText.slice(0, 2000)\",\n  targetId: \"task\",\n});\nreturn { url: read?.url, text: read?.result };\n```\n\nWhen evaluate is unavailable, keep the loop on structured state only.\n\n- Return only the fields the next step needs; never return the whole details object.\n- Completed cells share no state: re-embed the previous cell's returned `url`\u002F`newElements` in the next cell, or keep the comparison inside one cell. Only `waiting` runs persist, resumed by `wait`.\n- Interleave each act with a URL or tabs check before the next dependent act.\n- If a batch returns `aborted`, take a fresh snapshot before continuing.\n- If `newElements` is positive, inspect those elements first, then update the re-embedded state.\n- Use separate act calls when navigation is expected between steps.\n\n## Tab Hygiene\n\nBefore creating a tab for a named task, list tabs and reuse an existing matching label or URL when it is still usable.\n\nExample:\n\n```json\n{ \"action\": \"tabs\" }\n```\n\nIf no suitable tab exists:\n\n```json\n{ \"action\": \"open\", \"url\": \"https:\u002F\u002Fexample.com\", \"label\": \"task\" }\n```\n\nThen target it by label:\n\n```json\n{ \"action\": \"snapshot\", \"targetId\": \"task\", \"refs\": \"aria\" }\n```\n\nIf a retry creates duplicates, close the extras by `tabId`:\n\n```json\n{ \"action\": \"close\", \"targetId\": \"t3\" }\n```\n\nDo not pass bare numbers like `\"2\"` as `targetId`. Numeric tab positions are only for the CLI `openclaw browser tab select 2` helper; browser tool calls need a `suggestedTargetId`, label, `tabId`, or raw target id.\n\n## Stale Ref Recovery\n\nIf an action fails with a missing or stale ref:\n\n1. Snapshot the same `targetId` again.\n2. Find the current visible control.\n3. Retry once with the new ref.\n4. If the UI moved to a blocker state, report the blocker instead of looping.\n\n## Existing User Browser\n\nUse `profile=\"user\"` only when existing cookies\u002Flogin matter. This attaches to the user's running Chromium-based browser.\n\nOn macOS, `action=\"importprofile\"` is the alternative when the agent should use an isolated managed browser with cookies copied from a real Chrome-family profile. First use `action=\"profiles\"` and inspect `systemProfiles`, then import into a fresh managed profile name. Import asks for one Keychain\u002FTouch ID consent prompt. It copies cookies, not local storage or IndexedDB; device-bound session credentials (DBSC) mean some Google sessions may still require re-authentication.\n\nFor `profile=\"user\"` and other existing-session profiles, omit `timeoutMs` on `act:type`, `hover`, `scrollIntoView`, `drag`, `select`, and `fill`; that driver rejects per-call timeout overrides for those actions. `act:evaluate` accepts `timeoutMs`.\n\n## Google Meet Notes\n\nWhen creating or joining a Meet:\n\n- Treat camera\u002Fmicrophone permission screens as progress, not login failure.\n- If asked whether people can hear you, click the microphone option when voice is required.\n- If Google asks for sign-in, 2FA, account chooser confirmation, or permission that needs user approval, report the exact manual action.\n- Use one labeled tab per meeting flow, for example `label=\"meet\"`, and reuse it during retries.\n","# Browser Automation\n\nUse this skill when you need the `browser` tool for anything beyond a single page check.\n\n## Operating Loop\n\n1. Check browser state before acting:\n   - `openclaw browser doctor` or `action=\"status\"` when the browser\u002Fplugin setup itself may be broken.\n   - `action=\"status\"` for availability.\n   - `action=\"profiles\"` if login state or profile choice matters.\n   - `action=\"tabs\"` before opening a new tab if retries\u002Ftimeouts may have left windows behind.\n2. Prefer stable tab handles:\n   - Open important tabs with `label`, for example `label=\"meet\"`.\n   - After `action=\"tabs\"` or `action=\"open\"`, store `suggestedTargetId` and pass it as `targetId` in later calls.\n   - `suggestedTargetId` is the label when one exists, otherwise the stable `tabId` handle like `t1`.\n   - Avoid relying on raw DevTools `targetId` except for immediate diagnostics; it can change under Chromium target replacement.\n3. Read before you click:\n   - For “read the page and answer X,” use `action=\"text\"` with optional `selector` and `maxChars` for bounded visible prose (first selector match, otherwise article\u002Fmain\u002Fbody). On existing-session profiles, use `snapshot` instead. Efficient snapshots omit most prose.\n   - For virtualized lists, scroll through each segment, capture only the relevant rows, then merge the results.\n   - Use `action=\"snapshot\"` on the intended `targetId`.\n   - Add snapshot `query` to find lines containing all query tokens, ignoring case; matching lines keep their refs.\n   - Use the same `targetId` for follow-up actions so refs stay on the same tab.\n   - For durable Playwright refs, request `refs=\"aria\"` when supported. If you receive `axN` refs from `snapshotFormat=\"aria\"`, use them only after that same snapshot call; stale or unbound `axN` refs fail fast and need a fresh snapshot.\n   - Use `urls=true` when link text is ambiguous or a direct navigation target would avoid brittle clicks.\n   - Use `labels=true` on snapshot or screenshot when visual position matters. On Playwright-backed profiles, the response includes an `annotations` array (`{ref, number, role, name?, box}`) with each ref's bounding box in the captured image's coordinate space, so you can reason about position without re-snapshotting; screenshot labels can also combine with `fullPage=true` (CLI: `--full-page`) to label the whole document, or `ref` \u002F `element` to clip to one element. `profile=\"user\"` and other existing-session (chrome-mcp) profiles render an overlay into page screenshots but do not attach `annotations` or use the Playwright full-page\u002Fref\u002Felement projection helper, so read positions from the labeled image itself on those profiles. The raw-CDP fallback (no Playwright) does not support labeled screenshots at all and returns a 501, so only request `labels` when Playwright is available.\n4. Act narrowly:\n   - Prefer `action=\"act\"` with a ref from the latest snapshot.\n   - `navigate` returns the loaded page's compact snapshot inline, and batch `act` results that report a cross-document navigation include fresh page state; use those refs directly instead of a follow-up snapshot call.\n   - After a single act that triggers navigation, and after modal changes or form submissions, snapshot again before the next action.\n   - Avoid blind waits. Wait for visible UI state when possible.\n   - Use `action=\"emulate\"` with `device`, `colorScheme`, `timezoneId`, or `locale` when testing those settings; snapshot again afterward. Existing-session profiles do not support emulation.\n5. Report real blockers:\n   - Debug network failures with `action=\"requests\"`, optional URL\u002Ftype `filter`, and `limit` (default 50 recent entries). `clear=true` clears the collected log after reading. Use a managed profile; existing-session profiles do not support this log.\n   - Debug page errors with `action=\"errors\"` and `limit` (default 50 recent entries). `clear=true` clears the collected log after reading. Existing-session profiles do not support this log.\n   - If the page needs login, permission, captcha, 2FA, camera\u002Fmicrophone approval, or another manual step, stop and tell the user exactly what is needed.\n   - Do not claim the browser is not logged in just because the current page shows a permission or onboarding dialog. Inspect the visible UI first.\n\n## Browser batch CLI\n\n`openclaw browser batch` runs an array of nested `\u002Fact` actions in one `\u002Fact` call (the same `kind=\"batch\"` runtime reached through the agent tool), so CLI users and scripts can combine actions like `wait`, `click`, `type`, and `evaluate` into a single replayable plan without per-action round trips. Each entry in `actions[]` is a `BrowserActRequest` — the closed union the `\u002Fact` route accepts — not arbitrary `openclaw browser` subcommands. `batch` is not supported on `profile=\"user\"` and other existing-session (chrome-mcp) profiles; send actions individually there.\n\n- CLI: `openclaw browser batch --actions '\u003Cjson>'`, `--actions-file plan.json`, or `--actions-file -` for stdin. `--actions-file` and stdin input are capped at 1,000,000 bytes; split larger plans into multiple batch commands. `--continue` sets `stopOnError=false`; default stops on first error.\n- Ref lifecycle: refs come from a `snapshot` run before the batch (snapshot is not a nested action). A nested action that changes page state — such as a `click` that triggers navigation, or an `evaluate` that mutates the DOM — can invalidate earlier refs for the rest of the batch; put state-changing actions first, or split into a follow-up batch after re-snapshotting. Navigation and re-snapshotting happen outside the batch, since `open`, `navigate`, and `snapshot` are not `\u002Fact` kinds.\n- Target id: nested actions share the request's tab; an explicit nested `targetId` that resolves to a different tab is rejected with `ACT_TARGET_ID_MISMATCH`.\n- Response: `{ \"results\": [{ \"ok\": true } | { \"ok\": false, \"error\": \"...\" }, ...] }` in order; with default `stopOnError` the array ends at the first failure. Any failed entry exits nonzero; use `--json` to preserve the full response in scripts.\n\n## Code Mode Loop\n\nWhen `tools.codeMode` is enabled, the Browser tool has no normal turn — it is cataloged behind `exec`\u002F`wait`. Call it from exec cells as an async global, using the callable name the exec quick index advertises for the Browser tool (normally `browser`; colliding names get suffixed, and a client tool can win an identical name). An exact `catalog.search(\"browser\")` returns a handle already bound to the effective callable name, so resolve the handle in each cell and call it instead of hard-coding the literal global; an empty result means the Browser tool is not cataloged in this run.\n\nKeep the same labeled tab through the loop, and alternate reads with actions. Each `exec` cell starts a fresh VM — bindings from a completed cell are gone in the next, and only runs left `waiting` keep their state until `wait` resumes them — so carry comparison state across cells by returning it and re-embedding the returned values in the next cell:\n\n```javascript\n\u002F\u002F previous = the url\u002FnewElements returned by the last completed cell (a fresh\n\u002F\u002F VM runs this cell, so prior bindings do not exist here).\nconst previous = { url: \"https:\u002F\u002Fexample.com\u002Finbox\", newElements: 0 };\nconst [browser] = await catalog.search(\"browser\", { limit: 1 });\nconst details = await browser({\n  action: \"snapshot\",\n  snapshotFormat: \"ai\",\n  targetId: \"task\",\n  refs: \"aria\",\n  interactive: true,\n});\nconst changed =\n  details?.url !== previous.url ||\n  (details?.newElements ?? 0) > 0 ||\n  details?.blockedByDialog === true;\nreturn {\n  targetId: details?.targetId,\n  url: details?.url,\n  newElements: details?.newElements,\n  stats: details?.stats,\n  changed,\n};\n```\n\n- Code-mode calls return the tool's structured `details` directly (`targetId`, `url`, `newElements`, `stats`, `blockedByDialog`); rendered page text is not returned to code cells.\n- To read text inside code mode, run a targeted `act` evaluate (requires the evaluate capability; `browser.evaluateEnabled` can disable it) and keep the returned value bounded, because page-script output is untrusted:\n\n```javascript\nconst [browser] = await catalog.search(\"browser\", { limit: 1 });\nconst read = await browser({\n  action: \"act\",\n  kind: \"evaluate\",\n  fn: \"() => document.body.innerText.slice(0, 2000)\",\n  targetId: \"task\",\n});\nreturn { url: read?.url, text: read?.result };\n```\n\nWhen evaluate is unavailable, keep the loop on structured state only.\n\n- Return only the fields the next step needs; never return the whole details object.\n- Completed cells share no state: re-embed the previous cell's returned `url`\u002F`newElements` in the next cell, or keep the comparison inside one cell. Only `waiting` runs persist, resumed by `wait`.\n- Interleave each act with a URL or tabs check before the next dependent act.\n- If a batch returns `aborted`, take a fresh snapshot before continuing.\n- If `newElements` is positive, inspect those elements first, then update the re-embedded state.\n- Use separate act calls when navigation is expected between steps.\n\n## Tab Hygiene\n\nBefore creating a tab for a named task, list tabs and reuse an existing matching label or URL when it is still usable.\n\nExample:\n\n```json\n{ \"action\": \"tabs\" }\n```\n\nIf no suitable tab exists:\n\n```json\n{ \"action\": \"open\", \"url\": \"https:\u002F\u002Fexample.com\", \"label\": \"task\" }\n```\n\nThen target it by label:\n\n```json\n{ \"action\": \"snapshot\", \"targetId\": \"task\", \"refs\": \"aria\" }\n```\n\nIf a retry creates duplicates, close the extras by `tabId`:\n\n```json\n{ \"action\": \"close\", \"targetId\": \"t3\" }\n```\n\nDo not pass bare numbers like `\"2\"` as `targetId`. Numeric tab positions are only for the CLI `openclaw browser tab select 2` helper; browser tool calls need a `suggestedTargetId`, label, `tabId`, or raw target id.\n\n## Stale Ref Recovery\n\nIf an action fails with a missing or stale ref:\n\n1. Snapshot the same `targetId` again.\n2. Find the current visible control.\n3. Retry once with the new ref.\n4. If the UI moved to a blocker state, report the blocker instead of looping.\n\n## Existing User Browser\n\nUse `profile=\"user\"` only when existing cookies\u002Flogin matter. This attaches to the user's running Chromium-based browser.\n\nOn macOS, `action=\"importprofile\"` is the alternative when the agent should use an isolated managed browser with cookies copied from a real Chrome-family profile. First use `action=\"profiles\"` and inspect `systemProfiles`, then import into a fresh managed profile name. Import asks for one Keychain\u002FTouch ID consent prompt. It copies cookies, not local storage or IndexedDB; device-bound session credentials (DBSC) mean some Google sessions may still require re-authentication.\n\nFor `profile=\"user\"` and other existing-session profiles, omit `timeoutMs` on `act:type`, `hover`, `scrollIntoView`, `drag`, `select`, and `fill`; that driver rejects per-call timeout overrides for those actions. `act:evaluate` accepts `timeoutMs`.\n\n## Google Meet Notes\n\nWhen creating or joining a Meet:\n\n- Treat camera\u002Fmicrophone permission screens as progress, not login failure.\n- If asked whether people can hear you, click the microphone option when voice is required.\n- If Google asks for sign-in, 2FA, account chooser confirmation, or permission that needs user approval, report the exact manual action.\n- Use one labeled tab per meeting flow, for example `label=\"meet\"`, and reuse it during retries.\n",{"name":5,"description":7,"user-invocable":19},"main","54d89d355555ba29c51a05828bb2ee9b8abbfc1d","core\u002Fextensions\u002Fbrowser\u002Fskills\u002Fbrowser-automation","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI","https:\u002F\u002Fgithub.com\u002Fleon36000","2026-11-08 05:46:02",{"source":8,"skillName":5,"commands":30,"prompt":40,"zipAvailable":41},[31,34,37],{"tool":32,"command":33},"npx","npx skills add leon36000\u002FOPSYS-AI --skill browser-automation",{"tool":35,"command":36},"bunx","bunx skills add leon36000\u002FOPSYS-AI --skill browser-automation",{"tool":38,"command":39},"pnpm","pnpm dlx skills add leon36000\u002FOPSYS-AI --skill browser-automation","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Fcore\u002Fextensions\u002Fbrowser\u002Fskills\u002Fbrowser-automation\u002FSKILL.md",true,[],[44,51,58,65,72,79],{"skillNo":45,"name":46,"nameZh":6,"description":47,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":48,"categories":49,"keywords":50,"publishedAt":18},"01M79K66JYY28KMDVMJ4M1RBT7","procedure-demo","Procédure de test.","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Fdepartements\u002Fm09-demo\u002Fskills\u002Fprocedure-demo\u002FSKILL.md",[],[],{"skillNo":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":57,"publishedAt":18},"01M79K62M77K68EJYYT1HGZZYR","traiter-un-signal-critique","Procédure pour lister les signaux critiques d'une affaire et proposer leur prise en compte ou leur clôture, qui restent des décisions humaines, et pour proposer de lier une session à une affaire.","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Fdepartments\u002Fm10-memoire\u002Fskills\u002Ftraiter-un-signal-critique\u002FSKILL.md",[],[],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":18},"01M79K5XV1PC7Q699VC05WPVEX","preflight-contexte-suffisant","Procédure PREFLIGHT pour vérifier qu'un dossier de contexte (événement, résumé, preuves, impact, prochaine action, références de décisions) est suffisant avant une action importante, et pour consommer un Work Completion Trigger de M4.","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Fdepartments\u002Fm10-memoire\u002Fskills\u002Fpreflight-contexte-suffisant\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":18},"01M79K5TB07Q63XCV00FSX4AR0","journaliser-une-decision","Procédure pour journaliser une décision motivée au registre Affaires et obtenir sa référence vérifiable, ou pour signaler puis arbitrer une révision entre deux souvenirs, sans jamais fusionner ni muter un maillon existant.","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Fdepartments\u002Fm10-memoire\u002Fskills\u002Fjournaliser-une-decision\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":18},"01M79K5PB7JFBETB3A08ACKXXS","etat-consolide-et-contradictions","Procédure pour ingérer un snapshot d'un module, reconstruire l'état consolidé de l'entreprise et présenter ses contradictions internes sans jamais les résoudre ni en inventer la cause.","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Fdepartments\u002Fm09-strategic-intelligence\u002Fskills\u002Fetat-consolide-et-contradictions\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":18},"01M79K5K368MQEQA5TZ2FRQGMD","briefing-du-matin","Procédure pour produire le briefing du matin (décisions en attente, alertes M8, faits de la veille, manques) pour une date de référence donnée, lire ce qu'il n'a pas pu afficher, et lire les alertes de mesure de M8 avec leurs preuves, sans jamais rien y inventer, y décider ni y ajouter de diagnostic.","https:\u002F\u002Fgithub.com\u002Fleon36000\u002FOPSYS-AI\u002Fblob\u002Fmain\u002Fdepartments\u002Fm09-strategic-intelligence\u002Fskills\u002Fbriefing-du-matin\u002FSKILL.md",[],[]]