{ "schemaVersion": "1.0", "item": { "slug": "browserbase-sessions", "name": "browserbase-sessions", "source": "tencent", "type": "skill", "category": "开发工具", "sourceUrl": "https://clawhub.ai/JamesFincher/browserbase-sessions", "canonicalUrl": "https://clawhub.ai/JamesFincher/browserbase-sessions", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/browserbase-sessions", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=browserbase-sessions", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "pnpm-lock.yaml", "CHANGELOG.md", "references/api-quick-ref.md", "README.md", "package.json", "scripts/browserbase_manager.py" ], "primaryDoc": "SKILL.md", "quickSetup": [ "Download the package from Yavira.", "Extract the archive and review SKILL.md first.", "Import or place the package into your OpenClaw setup." ], "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run." } ] }, "sourceHealth": { "source": "tencent", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-05-07T17:22:31.273Z", "expiresAt": "2026-05-14T17:22:31.273Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-annual-report", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=afrexai-annual-report", "contentDisposition": "attachment; filename=\"afrexai-annual-report-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null }, "scope": "source", "summary": "Source download looks usable.", "detail": "Yavira can redirect you to the upstream package for this source.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/browserbase-sessions" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] }, "downloadPageUrl": "https://openagent3.xyz/downloads/browserbase-sessions", "agentPageUrl": "https://openagent3.xyz/skills/browserbase-sessions/agent", "manifestUrl": "https://openagent3.xyz/skills/browserbase-sessions/agent.json", "briefUrl": "https://openagent3.xyz/skills/browserbase-sessions/agent.md" }, "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run." } ] }, "documentation": { "source": "clawhub", "primaryDoc": "SKILL.md", "sections": [ { "title": "Browserbase Sessions Skill", "body": "Manage persistent cloud browser sessions via Browserbase. This skill creates browser sessions that preserve authentication (cookies, local storage) across interactions, automatically solve CAPTCHAs, and record sessions for later review." }, { "title": "Agent Checklist (Be Proactive)", "body": "If BROWSERBASE_API_KEY or BROWSERBASE_PROJECT_ID is missing, ask the user for them (and tell them where to find them). Do not run Browserbase commands until both are configured.\nIf commands fail due to missing Python deps (ImportError for browserbase / playwright), run:\n\npython3 {baseDir}/scripts/browserbase_manager.py install\nThen retry the original command.\n\n\nAsk the user what they want to persist and how they want to organize it:\n\nWorkspace per app/site (isolation): github, slack, stripe\nWorkspace per task/project (multi-site workflow): invoice-run, lead-gen, expense-recon\n\n\nWorkspaces persist:\n\nLogin state via Browserbase Contexts (cookies + storage)\nOpen tabs (URL + title snapshot) so you can restore where you left off\n\n\nPrefer workspace commands (create-workspace, start-workspace, resume-workspace, stop-workspace) over raw session commands when the user wants the browser to stay open across chat turns.\nPrefer direct interaction commands (list-tabs, new-tab, switch-tab, close-tab, click, type, press, wait-for, go-back, go-forward, reload, read-page) before falling back to execute-js.\nIf a workspace/session has a pending_handoff, check it first before doing anything else:\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff --action check --workspace \nIf not done, resend suggested_user_message and stop.\n\n\nWhenever a browser is opened (start-workspace, resume-workspace, or create-session), immediately share the human remote-control link:\n\nPrefer human_handoff.share_url from command output.\nPrefer human_handoff.share_text / human_handoff.share_markdown when replying to the user.\nFallback to human_control_url.\nIf missing, run live-url and share its human_handoff.share_url.\n\n\nWhen you need the user to perform a manual step (SSO/MFA/captcha/consent screens), use a handoff with a completion check:\n\nSet: python3 {baseDir}/scripts/browserbase_manager.py handoff --action set --workspace --instructions \"\" --url-contains \"\" (or --selector/--text/--cookie-name/...)\nVerify later: python3 {baseDir}/scripts/browserbase_manager.py handoff --action check --workspace (or --action wait)\n\n\nWhen closing, use stop-workspace (not terminate-session) so tabs are snapshotted and auth state is persisted." }, { "title": "Prompt-Optimized Response Patterns", "body": "Use short, consistent responses so users always know next actions.\n\nWhen credentials are missing:\n\nI need your Browserbase credentials before I can open a browser.\nPlease provide:\n1) BROWSERBASE_API_KEY\n2) BROWSERBASE_PROJECT_ID\n\nWhen browser is opened (session/workspace):\n\nBrowser is ready.\n\nI can keep working while you browse.\n\nWhen resuming an existing workspace:\n\nReconnected to your existing workspace.\n\n\nWhen a human step is required (SSO/MFA/consent screens):\n\nI need you to do one step in the live browser:\n1) \n\nOpen: \nStop when you reach: .\nI’ll detect it and continue automatically. If I don’t, reply “done” and I’ll re-check.\n\nWhen live URL is temporarily unavailable:\n\nThe remote-control URL is temporarily unavailable. I’ll retry now." }, { "title": "Step 1 — Get your Browserbase credentials", "body": "Sign up at browserbase.com if you haven't already.\nGo to Settings → API Keys and copy your API key (starts with bb_live_).\nGo to Settings → Project and copy your Project ID (a UUID).\n\nIf you have the API key but aren’t sure which Project ID to use, you can list projects:\n\nexport BROWSERBASE_API_KEY=\"bb_live_your_key_here\"\npython3 {baseDir}/scripts/browserbase_manager.py list-projects" }, { "title": "Step 2 — Install dependencies", "body": "Install Python deps and Playwright Chromium (recommended):\n\npython3 {baseDir}/scripts/browserbase_manager.py install\n\nManual alternative (pip/uv):\n\ncd {baseDir}/scripts && pip install -r requirements.txt\npython3 -m playwright install chromium" }, { "title": "Step 3 — Set environment variables", "body": "export BROWSERBASE_API_KEY=\"bb_live_your_key_here\"\nexport BROWSERBASE_PROJECT_ID=\"your-project-uuid-here\"\n\nOr configure via OpenClaw's skills.entries[\"browserbase-sessions\"].env in ~/.openclaw/openclaw.json (JSON5). Because this skill sets primaryEnv: BROWSERBASE_API_KEY, you can also use skills.entries[\"browserbase-sessions\"].apiKey for the API key:\n\n{\n skills: {\n entries: {\n \"browserbase-sessions\": {\n enabled: true,\n apiKey: \"bb_live_your_key_here\",\n env: {\n BROWSERBASE_PROJECT_ID: \"your-project-uuid-here\"\n }\n }\n }\n }\n}" }, { "title": "Step 4 — Run the setup test", "body": "This validates everything end-to-end (credentials, SDK, Playwright, API connection, and a live smoke test):\n\npython3 {baseDir}/scripts/browserbase_manager.py setup --install\n\nYou should see \"status\": \"success\" with all steps passing. If any step fails, the error message tells you exactly what to fix." }, { "title": "Defaults", "body": "Every session is created with these defaults to support research workflows:\n\nCaptcha solving: ON — Browserbase automatically solves CAPTCHAs so login flows and protected pages work without manual intervention. Disable with --no-solve-captchas.\nSession recording: ON — Browserbase records sessions (video in the Dashboard; rrweb events retrievable via API). Disable with --no-record.\nSession logging: ON — Browserbase captures session logs retrievable via API. Disable with --no-logs.\nAuth persistence — If you use a Context (or Workspace), auth state is persisted by default. Disable persistence with --no-persist." }, { "title": "Capabilities & Limitations (Be Explicit)", "body": "The agent can:\n\nCreate/inspect/terminate Browserbase sessions and contexts.\nKeep a browser “open” across chat turns using workspaces (keep-alive sessions + restore tabs).\nPersist login state across sessions via Browserbase Contexts (persist=true).\nRestore your place by reopening the last saved set of open tabs (URL + title snapshot).\nProvide a live debugger URL so the user can browse manually while the agent continues working.\nUse interactive browser controls: list/open/switch/close tabs, click/type/press keys, wait for selectors/text/url states, go back/forward/reload, and read page text/html/links.\nTake screenshots, run JavaScript, read cookies, fetch logs, and fetch rrweb recording events.\n\nThe agent cannot:\n\nKeep sessions running indefinitely (Browserbase enforces timeouts; max is 6 hours).\nRestore full back/forward browser history (only open URLs are restored).\nPassively “watch” what the user is doing in the live debugger. To detect human actions, the agent must reconnect and check for specific completion conditions (selector/text/url/cookie/storage) using handoff or wait-for.\nBypass MFA/SSO without user participation.\nDownload the Dashboard video via API (the API returns rrweb events, not a video file)." }, { "title": "Available Commands", "body": "All commands are run via the manager script:\n\npython3 {baseDir}/scripts/browserbase_manager.py [options]" }, { "title": "Setup & Validation", "body": "Install deps (only needed once per environment):\n\npython3 {baseDir}/scripts/browserbase_manager.py install\n\nRun the full setup test:\n\npython3 {baseDir}/scripts/browserbase_manager.py setup --install" }, { "title": "Workspaces (Recommended)", "body": "Workspaces are the recommended way to keep a browser \"open\" while chatting and then pick up later. A workspace combines:\n\nA Browserbase Context (persists cookies + local/session storage, so you stay logged in)\nA local tab snapshot (URLs + titles) so tabs can be restored into the next session (note: this restores open URLs, not full back/forward browser history)\nThe current active session id so the agent can reconnect\n\nTask workspaces (multi-site flows)\n\nA single Browserbase Context is a browser profile, so it can keep you logged into multiple sites at once. For workflows like “do something on Site A, then do something on Site B”, create a task workspace and keep both sites open as tabs:\n\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name invoice-run\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name invoice-run --timeout 21600\npython3 {baseDir}/scripts/browserbase_manager.py live-url --workspace invoice-run\n\nIf you need account/cookie isolation (different logins, fewer cross-site side effects), use separate workspaces per app/site instead.\n\nCreate and start a workspace:\n\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name github\npython3 {baseDir}/scripts/browserbase_manager.py list-workspaces\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name github --timeout 21600\n# Share this field with the user immediately:\n# human_handoff.share_url (fallback: human_control_url / live_urls.debugger_url)\n\nNote: start-workspace performs a short “warm connect” via Playwright so the session doesn’t die from the 5-minute connect requirement, even if the user hasn’t opened the live debugger yet.\n\nWhile the user is browsing in the live debugger, the agent can keep working. To resume later:\n\npython3 {baseDir}/scripts/browserbase_manager.py resume-workspace --name github\n\nFor long-running sessions (especially when the user is opening/closing tabs manually), take snapshots periodically:\n\npython3 {baseDir}/scripts/browserbase_manager.py snapshot-workspace --name github\n\nTo persist login + tabs when you are done, always stop via the workspace command:\n\npython3 {baseDir}/scripts/browserbase_manager.py stop-workspace --name github\n\nTo inspect what a workspace has saved (context id, active session id, tabs, history):\n\npython3 {baseDir}/scripts/browserbase_manager.py get-workspace --name github\n\nMost commands accept --workspace instead of --session-id:\n\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace github --url \"https://github.com/settings/profile\"\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --workspace github --output /tmp/profile.png\npython3 {baseDir}/scripts/browserbase_manager.py execute-js --workspace github --code \"document.title\"" }, { "title": "Context Management (for authentication persistence)", "body": "Create a named context to store login state:\n\npython3 {baseDir}/scripts/browserbase_manager.py create-context --name github\n\nList all saved contexts:\n\npython3 {baseDir}/scripts/browserbase_manager.py list-contexts\n\nDelete a context (by name or ID):\n\npython3 {baseDir}/scripts/browserbase_manager.py delete-context --context-id github" }, { "title": "Session Lifecycle", "body": "Create a new session (captcha solving, recording, and logging enabled by default):\n\n# Basic session\npython3 {baseDir}/scripts/browserbase_manager.py create-session\n\n# Session with saved context (persist=true by default when a context is used)\npython3 {baseDir}/scripts/browserbase_manager.py create-session --context-id github\n\n# Keep-alive session for long research (survives disconnections)\npython3 {baseDir}/scripts/browserbase_manager.py create-session --context-id github --keep-alive --timeout 3600\n\n# Full options\npython3 {baseDir}/scripts/browserbase_manager.py create-session \\\n --context-id github \\\n --keep-alive \\\n --timeout 3600 \\\n --region us-west-2 \\\n --proxy \\\n --block-ads \\\n --viewport-width 1280 \\\n --viewport-height 720\n\nList all sessions:\n\npython3 {baseDir}/scripts/browserbase_manager.py list-sessions\npython3 {baseDir}/scripts/browserbase_manager.py list-sessions --status RUNNING\n\nGet session details:\n\npython3 {baseDir}/scripts/browserbase_manager.py get-session --session-id \n\nTerminate a session:\n\npython3 {baseDir}/scripts/browserbase_manager.py terminate-session --session-id " }, { "title": "Browser Automation", "body": "Navigate to a URL:\n\n# Navigate and get page title\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\"\n\n# Navigate and extract text\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\" --extract-text\n\n# Navigate and save screenshot\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\" --screenshot /tmp/page.png\n\n# Navigate and take full-page screenshot\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\" --screenshot /tmp/full.png --full-page\n\nManage tabs:\n\npython3 {baseDir}/scripts/browserbase_manager.py list-tabs --session-id \npython3 {baseDir}/scripts/browserbase_manager.py new-tab --session-id --url \"https://example.org\"\npython3 {baseDir}/scripts/browserbase_manager.py switch-tab --session-id --tab-index 1\npython3 {baseDir}/scripts/browserbase_manager.py close-tab --session-id --tab-url-contains \"example.org\"\n\nInteract with the page:\n\npython3 {baseDir}/scripts/browserbase_manager.py click --session-id --selector \"button[type='submit']\"\npython3 {baseDir}/scripts/browserbase_manager.py type --session-id --selector \"input[name='email']\" --text \"user@example.com\" --clear\npython3 {baseDir}/scripts/browserbase_manager.py press --session-id --key \"Enter\"\npython3 {baseDir}/scripts/browserbase_manager.py wait-for --session-id --selector \".dashboard-ready\" --timeout-ms 45000\n\nControl navigation state:\n\npython3 {baseDir}/scripts/browserbase_manager.py go-back --session-id \npython3 {baseDir}/scripts/browserbase_manager.py go-forward --session-id \npython3 {baseDir}/scripts/browserbase_manager.py reload --session-id \n\nRead the current page:\n\npython3 {baseDir}/scripts/browserbase_manager.py read-page --session-id --max-text-chars 20000\npython3 {baseDir}/scripts/browserbase_manager.py read-page --session-id --include-links --max-links 30\npython3 {baseDir}/scripts/browserbase_manager.py read-page --session-id --include-html --max-html-chars 120000\n\nTake a screenshot of the current page (without navigating):\n\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --session-id --output /tmp/current.png\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --session-id --output /tmp/full.png --full-page\n\nExecute JavaScript:\n\npython3 {baseDir}/scripts/browserbase_manager.py execute-js --session-id --code \"document.title\"\n\nGet cookies:\n\npython3 {baseDir}/scripts/browserbase_manager.py get-cookies --session-id \n\nAll of the commands above also support --workspace so the active workspace session is used automatically." }, { "title": "Recordings, Logs & Debug", "body": "Fetch rrweb recording events (session must be terminated first):\n\npython3 {baseDir}/scripts/browserbase_manager.py get-recording --session-id --output /tmp/session.rrweb.json\n\nDownload files saved during the session (archive):\n\npython3 {baseDir}/scripts/browserbase_manager.py get-downloads --session-id --output /tmp/downloads.zip\n\nGet session logs:\n\npython3 {baseDir}/scripts/browserbase_manager.py get-logs --session-id \n\nGet the live debug URL (for visual inspection of a running session):\n\npython3 {baseDir}/scripts/browserbase_manager.py live-url --session-id \n# Share: human_handoff.share_url" }, { "title": "Human Handoff (Manual Steps)", "body": "Use handoff when the user must do something manually (SSO/MFA/consent screens) and you want to detect completion and resume automatically.\n\nCompletion checks (pick the most reliable signal you can):\n\nPrefer --url-contains for post-login destinations (e.g. /dashboard, /app, /settings).\nUse --selector when the URL doesn’t change (SPAs).\nUse --cookie-name, --local-storage-key, --session-storage-key only when you know a stable auth indicator.\n\nCombining checks:\n\nDefault is --match all (AND): all provided checks must be true.\nUse --match any (OR) when you want multiple fallback signals (e.g., URL contains /dashboard OR selector .dashboard-ready).\n\nSet a handoff (store the instructions + completion check):\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff \\\n --action set \\\n --workspace myapp \\\n --instructions \"Log in and stop on the dashboard.\" \\\n --url-contains \"/dashboard\"\n\nFallback example (any-of):\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff \\\n --action set \\\n --workspace myapp \\\n --instructions \"Complete login and stop on the dashboard.\" \\\n --url-contains \"/dashboard\" \\\n --selector \".dashboard-ready\" \\\n --match any\n\nTip: handoff --action set returns suggested_user_message.text / suggested_user_message.markdown so you can paste a consistent “hey human do this” message (with the live debugger URL) to the user.\n\nLater, verify once:\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff --action check --workspace myapp\n\nOr wait up to 5 minutes (default) for completion:\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff --action wait --workspace myapp" }, { "title": "Workflow 1: Multi-session research with persistent login", "body": "# 1. One-time: create a workspace for the site (creates a Browserbase Context + local state)\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name myapp\n\n# 2. Start a keep-alive session (tabs restored from last snapshot, login persisted via context)\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name myapp --timeout 3600\n\n# 3. Open the live debugger URL so the user can log in / browse while you keep chatting\npython3 {baseDir}/scripts/browserbase_manager.py live-url --workspace myapp\n\n# 4. Do research, take screenshots (workspace auto-tracks tabs + history)\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace myapp --url \"https://myapp.com/dashboard\" --extract-text\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --workspace myapp --output /tmp/dashboard.png\n\n# 5. When done: stop-workspace snapshots tabs + persists auth state back to the context\npython3 {baseDir}/scripts/browserbase_manager.py stop-workspace --name myapp\n\n# 6. Later: resume and pick up where you left off\npython3 {baseDir}/scripts/browserbase_manager.py resume-workspace --name myapp" }, { "title": "Workflow 1b: Task workflow across multiple sites (persist tabs + logins)", "body": "# 1) Create a task workspace (one browser profile that can stay logged into multiple sites)\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name lead-gen\n\n# 2) Start it and open the live debugger so the user can log in on both sites\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name lead-gen --timeout 21600\npython3 {baseDir}/scripts/browserbase_manager.py live-url --workspace lead-gen\n\n# 3) Agent can navigate and capture state while you chat (tabs snapshot includes both Site A and Site B)\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace lead-gen --url \"https://site-a.example.com\" --extract-text\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace lead-gen --url \"https://site-b.example.com\" --extract-text\n\n# 4) If the user is manually opening/closing tabs in the debugger, snapshot occasionally:\npython3 {baseDir}/scripts/browserbase_manager.py snapshot-workspace --name lead-gen\n\n# 5) Stop to persist auth + tabs snapshot\npython3 {baseDir}/scripts/browserbase_manager.py stop-workspace --name lead-gen" }, { "title": "Workflow 2: Screenshot documentation", "body": "python3 {baseDir}/scripts/browserbase_manager.py create-session\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://docs.example.com\" --screenshot /tmp/docs_home.png\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://docs.example.com/api\" --screenshot /tmp/docs_api.png --full-page\npython3 {baseDir}/scripts/browserbase_manager.py terminate-session --session-id " }, { "title": "Workflow 3: Record and share a walkthrough", "body": "# Session recording is ON by default\npython3 {baseDir}/scripts/browserbase_manager.py create-session --context-id myapp\n# ... do your walkthrough (navigate, click, etc.) ...\npython3 {baseDir}/scripts/browserbase_manager.py terminate-session --session-id \n# Save rrweb recording events\n# (Video is available in the Browserbase Dashboard; this fetches rrweb events)\npython3 {baseDir}/scripts/browserbase_manager.py get-recording --session-id --output /tmp/walkthrough.rrweb.json" }, { "title": "Important Notes", "body": "Captcha solving is ON by default. Browserbase handles CAPTCHAs automatically during login flows and page loads. Use --no-solve-captchas to disable.\nRecording is ON by default. Video is available in the Browserbase Dashboard; get-recording fetches rrweb events (primary tab) for programmatic replay. Use --no-record to disable.\nLogging is ON by default. Use get-logs to retrieve logs; disable with --no-logs.\nConnection timeout: 5 minutes to connect after creation before auto-termination.\nKeep-alive sessions survive disconnections and must be explicitly terminated.\nContext persistence: When a session was created with a context using persist=true (default), wait a few seconds after termination before creating a new session with the same context.\nNamed contexts: Use --name with create-context to save friendly names (e.g. github, slack). Use the name anywhere a context ID is expected.\nWorkspace state: Workspaces are stored locally under ~/.browserbase/workspaces/.json (or BROWSERBASE_CONFIG_DIR/workspaces). They include the context id, active session id, and the last saved tab snapshot.\nOne context per site: Use separate contexts for different authenticated sites.\nAvoid concurrent sessions on the same context.\nRegions: us-west-2 (default), us-east-1, eu-central-1, ap-southeast-1.\nSession timeout: 60–21600 seconds (max 6 hours).\nCosts/limits: Your Browserbase plan has limits (browser hours, proxy data, concurrency). Keep-alive sessions consume hours while running; terminate sessions and set reasonable --timeout values to control cost. Check your Browserbase dashboard for current quotas." }, { "title": "Error Handling", "body": "All commands return JSON output. On error, the output includes an \"error\" key. Common errors:\n\nAPIConnectionError: Browserbase API unreachable\nRateLimitError: Too many concurrent sessions for your plan\nAPIStatusError: Invalid parameters or authentication failure\nMissing env vars: Set BROWSERBASE_API_KEY and BROWSERBASE_PROJECT_ID" }, { "title": "Reference", "body": "For full API details, read {baseDir}/references/api-quick-ref.md." } ], "body": "Browserbase Sessions Skill\n\nManage persistent cloud browser sessions via Browserbase. This skill creates browser sessions that preserve authentication (cookies, local storage) across interactions, automatically solve CAPTCHAs, and record sessions for later review.\n\nAgent Checklist (Be Proactive)\nIf BROWSERBASE_API_KEY or BROWSERBASE_PROJECT_ID is missing, ask the user for them (and tell them where to find them). Do not run Browserbase commands until both are configured.\nIf commands fail due to missing Python deps (ImportError for browserbase / playwright), run:\npython3 {baseDir}/scripts/browserbase_manager.py install\nThen retry the original command.\nAsk the user what they want to persist and how they want to organize it:\nWorkspace per app/site (isolation): github, slack, stripe\nWorkspace per task/project (multi-site workflow): invoice-run, lead-gen, expense-recon\nWorkspaces persist:\nLogin state via Browserbase Contexts (cookies + storage)\nOpen tabs (URL + title snapshot) so you can restore where you left off\nPrefer workspace commands (create-workspace, start-workspace, resume-workspace, stop-workspace) over raw session commands when the user wants the browser to stay open across chat turns.\nPrefer direct interaction commands (list-tabs, new-tab, switch-tab, close-tab, click, type, press, wait-for, go-back, go-forward, reload, read-page) before falling back to execute-js.\nIf a workspace/session has a pending_handoff, check it first before doing anything else:\npython3 {baseDir}/scripts/browserbase_manager.py handoff --action check --workspace \nIf not done, resend suggested_user_message and stop.\nWhenever a browser is opened (start-workspace, resume-workspace, or create-session), immediately share the human remote-control link:\nPrefer human_handoff.share_url from command output.\nPrefer human_handoff.share_text / human_handoff.share_markdown when replying to the user.\nFallback to human_control_url.\nIf missing, run live-url and share its human_handoff.share_url.\nWhen you need the user to perform a manual step (SSO/MFA/captcha/consent screens), use a handoff with a completion check:\nSet: python3 {baseDir}/scripts/browserbase_manager.py handoff --action set --workspace --instructions \"\" --url-contains \"\" (or --selector/--text/--cookie-name/...)\nVerify later: python3 {baseDir}/scripts/browserbase_manager.py handoff --action check --workspace (or --action wait)\nWhen closing, use stop-workspace (not terminate-session) so tabs are snapshotted and auth state is persisted.\nPrompt-Optimized Response Patterns\n\nUse short, consistent responses so users always know next actions.\n\nWhen credentials are missing:\n\nI need your Browserbase credentials before I can open a browser.\nPlease provide:\n1) BROWSERBASE_API_KEY\n2) BROWSERBASE_PROJECT_ID\n\n\nWhen browser is opened (session/workspace):\n\nBrowser is ready.\n\nI can keep working while you browse.\n\n\nWhen resuming an existing workspace:\n\nReconnected to your existing workspace.\n\n\n\nWhen a human step is required (SSO/MFA/consent screens):\n\nI need you to do one step in the live browser:\n1) \n\nOpen: \nStop when you reach: .\nI’ll detect it and continue automatically. If I don’t, reply “done” and I’ll re-check.\n\n\nWhen live URL is temporarily unavailable:\n\nThe remote-control URL is temporarily unavailable. I’ll retry now.\n\nFirst-Time Setup\nStep 1 — Get your Browserbase credentials\nSign up at browserbase.com if you haven't already.\nGo to Settings → API Keys and copy your API key (starts with bb_live_).\nGo to Settings → Project and copy your Project ID (a UUID).\n\nIf you have the API key but aren’t sure which Project ID to use, you can list projects:\n\nexport BROWSERBASE_API_KEY=\"bb_live_your_key_here\"\npython3 {baseDir}/scripts/browserbase_manager.py list-projects\n\nStep 2 — Install dependencies\n\nInstall Python deps and Playwright Chromium (recommended):\n\npython3 {baseDir}/scripts/browserbase_manager.py install\n\n\nManual alternative (pip/uv):\n\ncd {baseDir}/scripts && pip install -r requirements.txt\npython3 -m playwright install chromium\n\nStep 3 — Set environment variables\nexport BROWSERBASE_API_KEY=\"bb_live_your_key_here\"\nexport BROWSERBASE_PROJECT_ID=\"your-project-uuid-here\"\n\n\nOr configure via OpenClaw's skills.entries[\"browserbase-sessions\"].env in ~/.openclaw/openclaw.json (JSON5). Because this skill sets primaryEnv: BROWSERBASE_API_KEY, you can also use skills.entries[\"browserbase-sessions\"].apiKey for the API key:\n\n{\n skills: {\n entries: {\n \"browserbase-sessions\": {\n enabled: true,\n apiKey: \"bb_live_your_key_here\",\n env: {\n BROWSERBASE_PROJECT_ID: \"your-project-uuid-here\"\n }\n }\n }\n }\n}\n\nStep 4 — Run the setup test\n\nThis validates everything end-to-end (credentials, SDK, Playwright, API connection, and a live smoke test):\n\npython3 {baseDir}/scripts/browserbase_manager.py setup --install\n\n\nYou should see \"status\": \"success\" with all steps passing. If any step fails, the error message tells you exactly what to fix.\n\nDefaults\n\nEvery session is created with these defaults to support research workflows:\n\nCaptcha solving: ON — Browserbase automatically solves CAPTCHAs so login flows and protected pages work without manual intervention. Disable with --no-solve-captchas.\nSession recording: ON — Browserbase records sessions (video in the Dashboard; rrweb events retrievable via API). Disable with --no-record.\nSession logging: ON — Browserbase captures session logs retrievable via API. Disable with --no-logs.\nAuth persistence — If you use a Context (or Workspace), auth state is persisted by default. Disable persistence with --no-persist.\nCapabilities & Limitations (Be Explicit)\n\nThe agent can:\n\nCreate/inspect/terminate Browserbase sessions and contexts.\nKeep a browser “open” across chat turns using workspaces (keep-alive sessions + restore tabs).\nPersist login state across sessions via Browserbase Contexts (persist=true).\nRestore your place by reopening the last saved set of open tabs (URL + title snapshot).\nProvide a live debugger URL so the user can browse manually while the agent continues working.\nUse interactive browser controls: list/open/switch/close tabs, click/type/press keys, wait for selectors/text/url states, go back/forward/reload, and read page text/html/links.\nTake screenshots, run JavaScript, read cookies, fetch logs, and fetch rrweb recording events.\n\nThe agent cannot:\n\nKeep sessions running indefinitely (Browserbase enforces timeouts; max is 6 hours).\nRestore full back/forward browser history (only open URLs are restored).\nPassively “watch” what the user is doing in the live debugger. To detect human actions, the agent must reconnect and check for specific completion conditions (selector/text/url/cookie/storage) using handoff or wait-for.\nBypass MFA/SSO without user participation.\nDownload the Dashboard video via API (the API returns rrweb events, not a video file).\nAvailable Commands\n\nAll commands are run via the manager script:\n\npython3 {baseDir}/scripts/browserbase_manager.py [options]\n\nSetup & Validation\n\nInstall deps (only needed once per environment):\n\npython3 {baseDir}/scripts/browserbase_manager.py install\n\n\nRun the full setup test:\n\npython3 {baseDir}/scripts/browserbase_manager.py setup --install\n\nWorkspaces (Recommended)\n\nWorkspaces are the recommended way to keep a browser \"open\" while chatting and then pick up later. A workspace combines:\n\nA Browserbase Context (persists cookies + local/session storage, so you stay logged in)\nA local tab snapshot (URLs + titles) so tabs can be restored into the next session (note: this restores open URLs, not full back/forward browser history)\nThe current active session id so the agent can reconnect\nTask workspaces (multi-site flows)\n\nA single Browserbase Context is a browser profile, so it can keep you logged into multiple sites at once. For workflows like “do something on Site A, then do something on Site B”, create a task workspace and keep both sites open as tabs:\n\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name invoice-run\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name invoice-run --timeout 21600\npython3 {baseDir}/scripts/browserbase_manager.py live-url --workspace invoice-run\n\n\nIf you need account/cookie isolation (different logins, fewer cross-site side effects), use separate workspaces per app/site instead.\n\nCreate and start a workspace:\n\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name github\npython3 {baseDir}/scripts/browserbase_manager.py list-workspaces\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name github --timeout 21600\n# Share this field with the user immediately:\n# human_handoff.share_url (fallback: human_control_url / live_urls.debugger_url)\n\n\nNote: start-workspace performs a short “warm connect” via Playwright so the session doesn’t die from the 5-minute connect requirement, even if the user hasn’t opened the live debugger yet.\n\nWhile the user is browsing in the live debugger, the agent can keep working. To resume later:\n\npython3 {baseDir}/scripts/browserbase_manager.py resume-workspace --name github\n\n\nFor long-running sessions (especially when the user is opening/closing tabs manually), take snapshots periodically:\n\npython3 {baseDir}/scripts/browserbase_manager.py snapshot-workspace --name github\n\n\nTo persist login + tabs when you are done, always stop via the workspace command:\n\npython3 {baseDir}/scripts/browserbase_manager.py stop-workspace --name github\n\n\nTo inspect what a workspace has saved (context id, active session id, tabs, history):\n\npython3 {baseDir}/scripts/browserbase_manager.py get-workspace --name github\n\n\nMost commands accept --workspace instead of --session-id:\n\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace github --url \"https://github.com/settings/profile\"\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --workspace github --output /tmp/profile.png\npython3 {baseDir}/scripts/browserbase_manager.py execute-js --workspace github --code \"document.title\"\n\nContext Management (for authentication persistence)\n\nCreate a named context to store login state:\n\npython3 {baseDir}/scripts/browserbase_manager.py create-context --name github\n\n\nList all saved contexts:\n\npython3 {baseDir}/scripts/browserbase_manager.py list-contexts\n\n\nDelete a context (by name or ID):\n\npython3 {baseDir}/scripts/browserbase_manager.py delete-context --context-id github\n\nSession Lifecycle\n\nCreate a new session (captcha solving, recording, and logging enabled by default):\n\n# Basic session\npython3 {baseDir}/scripts/browserbase_manager.py create-session\n\n# Session with saved context (persist=true by default when a context is used)\npython3 {baseDir}/scripts/browserbase_manager.py create-session --context-id github\n\n# Keep-alive session for long research (survives disconnections)\npython3 {baseDir}/scripts/browserbase_manager.py create-session --context-id github --keep-alive --timeout 3600\n\n# Full options\npython3 {baseDir}/scripts/browserbase_manager.py create-session \\\n --context-id github \\\n --keep-alive \\\n --timeout 3600 \\\n --region us-west-2 \\\n --proxy \\\n --block-ads \\\n --viewport-width 1280 \\\n --viewport-height 720\n\n\nList all sessions:\n\npython3 {baseDir}/scripts/browserbase_manager.py list-sessions\npython3 {baseDir}/scripts/browserbase_manager.py list-sessions --status RUNNING\n\n\nGet session details:\n\npython3 {baseDir}/scripts/browserbase_manager.py get-session --session-id \n\n\nTerminate a session:\n\npython3 {baseDir}/scripts/browserbase_manager.py terminate-session --session-id \n\nBrowser Automation\n\nNavigate to a URL:\n\n# Navigate and get page title\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\"\n\n# Navigate and extract text\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\" --extract-text\n\n# Navigate and save screenshot\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\" --screenshot /tmp/page.png\n\n# Navigate and take full-page screenshot\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://example.com\" --screenshot /tmp/full.png --full-page\n\n\nManage tabs:\n\npython3 {baseDir}/scripts/browserbase_manager.py list-tabs --session-id \npython3 {baseDir}/scripts/browserbase_manager.py new-tab --session-id --url \"https://example.org\"\npython3 {baseDir}/scripts/browserbase_manager.py switch-tab --session-id --tab-index 1\npython3 {baseDir}/scripts/browserbase_manager.py close-tab --session-id --tab-url-contains \"example.org\"\n\n\nInteract with the page:\n\npython3 {baseDir}/scripts/browserbase_manager.py click --session-id --selector \"button[type='submit']\"\npython3 {baseDir}/scripts/browserbase_manager.py type --session-id --selector \"input[name='email']\" --text \"user@example.com\" --clear\npython3 {baseDir}/scripts/browserbase_manager.py press --session-id --key \"Enter\"\npython3 {baseDir}/scripts/browserbase_manager.py wait-for --session-id --selector \".dashboard-ready\" --timeout-ms 45000\n\n\nControl navigation state:\n\npython3 {baseDir}/scripts/browserbase_manager.py go-back --session-id \npython3 {baseDir}/scripts/browserbase_manager.py go-forward --session-id \npython3 {baseDir}/scripts/browserbase_manager.py reload --session-id \n\n\nRead the current page:\n\npython3 {baseDir}/scripts/browserbase_manager.py read-page --session-id --max-text-chars 20000\npython3 {baseDir}/scripts/browserbase_manager.py read-page --session-id --include-links --max-links 30\npython3 {baseDir}/scripts/browserbase_manager.py read-page --session-id --include-html --max-html-chars 120000\n\n\nTake a screenshot of the current page (without navigating):\n\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --session-id --output /tmp/current.png\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --session-id --output /tmp/full.png --full-page\n\n\nExecute JavaScript:\n\npython3 {baseDir}/scripts/browserbase_manager.py execute-js --session-id --code \"document.title\"\n\n\nGet cookies:\n\npython3 {baseDir}/scripts/browserbase_manager.py get-cookies --session-id \n\n\nAll of the commands above also support --workspace so the active workspace session is used automatically.\n\nRecordings, Logs & Debug\n\nFetch rrweb recording events (session must be terminated first):\n\npython3 {baseDir}/scripts/browserbase_manager.py get-recording --session-id --output /tmp/session.rrweb.json\n\n\nDownload files saved during the session (archive):\n\npython3 {baseDir}/scripts/browserbase_manager.py get-downloads --session-id --output /tmp/downloads.zip\n\n\nGet session logs:\n\npython3 {baseDir}/scripts/browserbase_manager.py get-logs --session-id \n\n\nGet the live debug URL (for visual inspection of a running session):\n\npython3 {baseDir}/scripts/browserbase_manager.py live-url --session-id \n# Share: human_handoff.share_url\n\nHuman Handoff (Manual Steps)\n\nUse handoff when the user must do something manually (SSO/MFA/consent screens) and you want to detect completion and resume automatically.\n\nCompletion checks (pick the most reliable signal you can):\n\nPrefer --url-contains for post-login destinations (e.g. /dashboard, /app, /settings).\nUse --selector when the URL doesn’t change (SPAs).\nUse --cookie-name, --local-storage-key, --session-storage-key only when you know a stable auth indicator.\n\nCombining checks:\n\nDefault is --match all (AND): all provided checks must be true.\nUse --match any (OR) when you want multiple fallback signals (e.g., URL contains /dashboard OR selector .dashboard-ready).\n\nSet a handoff (store the instructions + completion check):\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff \\\n --action set \\\n --workspace myapp \\\n --instructions \"Log in and stop on the dashboard.\" \\\n --url-contains \"/dashboard\"\n\n\nFallback example (any-of):\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff \\\n --action set \\\n --workspace myapp \\\n --instructions \"Complete login and stop on the dashboard.\" \\\n --url-contains \"/dashboard\" \\\n --selector \".dashboard-ready\" \\\n --match any\n\n\nTip: handoff --action set returns suggested_user_message.text / suggested_user_message.markdown so you can paste a consistent “hey human do this” message (with the live debugger URL) to the user.\n\nLater, verify once:\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff --action check --workspace myapp\n\n\nOr wait up to 5 minutes (default) for completion:\n\npython3 {baseDir}/scripts/browserbase_manager.py handoff --action wait --workspace myapp\n\nCommon Workflows\nWorkflow 1: Multi-session research with persistent login\n# 1. One-time: create a workspace for the site (creates a Browserbase Context + local state)\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name myapp\n\n# 2. Start a keep-alive session (tabs restored from last snapshot, login persisted via context)\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name myapp --timeout 3600\n\n# 3. Open the live debugger URL so the user can log in / browse while you keep chatting\npython3 {baseDir}/scripts/browserbase_manager.py live-url --workspace myapp\n\n# 4. Do research, take screenshots (workspace auto-tracks tabs + history)\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace myapp --url \"https://myapp.com/dashboard\" --extract-text\npython3 {baseDir}/scripts/browserbase_manager.py screenshot --workspace myapp --output /tmp/dashboard.png\n\n# 5. When done: stop-workspace snapshots tabs + persists auth state back to the context\npython3 {baseDir}/scripts/browserbase_manager.py stop-workspace --name myapp\n\n# 6. Later: resume and pick up where you left off\npython3 {baseDir}/scripts/browserbase_manager.py resume-workspace --name myapp\n\nWorkflow 1b: Task workflow across multiple sites (persist tabs + logins)\n# 1) Create a task workspace (one browser profile that can stay logged into multiple sites)\npython3 {baseDir}/scripts/browserbase_manager.py create-workspace --name lead-gen\n\n# 2) Start it and open the live debugger so the user can log in on both sites\npython3 {baseDir}/scripts/browserbase_manager.py start-workspace --name lead-gen --timeout 21600\npython3 {baseDir}/scripts/browserbase_manager.py live-url --workspace lead-gen\n\n# 3) Agent can navigate and capture state while you chat (tabs snapshot includes both Site A and Site B)\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace lead-gen --url \"https://site-a.example.com\" --extract-text\npython3 {baseDir}/scripts/browserbase_manager.py navigate --workspace lead-gen --url \"https://site-b.example.com\" --extract-text\n\n# 4) If the user is manually opening/closing tabs in the debugger, snapshot occasionally:\npython3 {baseDir}/scripts/browserbase_manager.py snapshot-workspace --name lead-gen\n\n# 5) Stop to persist auth + tabs snapshot\npython3 {baseDir}/scripts/browserbase_manager.py stop-workspace --name lead-gen\n\nWorkflow 2: Screenshot documentation\npython3 {baseDir}/scripts/browserbase_manager.py create-session\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://docs.example.com\" --screenshot /tmp/docs_home.png\npython3 {baseDir}/scripts/browserbase_manager.py navigate --session-id --url \"https://docs.example.com/api\" --screenshot /tmp/docs_api.png --full-page\npython3 {baseDir}/scripts/browserbase_manager.py terminate-session --session-id \n\nWorkflow 3: Record and share a walkthrough\n# Session recording is ON by default\npython3 {baseDir}/scripts/browserbase_manager.py create-session --context-id myapp\n# ... do your walkthrough (navigate, click, etc.) ...\npython3 {baseDir}/scripts/browserbase_manager.py terminate-session --session-id \n# Save rrweb recording events\n# (Video is available in the Browserbase Dashboard; this fetches rrweb events)\npython3 {baseDir}/scripts/browserbase_manager.py get-recording --session-id --output /tmp/walkthrough.rrweb.json\n\nImportant Notes\nCaptcha solving is ON by default. Browserbase handles CAPTCHAs automatically during login flows and page loads. Use --no-solve-captchas to disable.\nRecording is ON by default. Video is available in the Browserbase Dashboard; get-recording fetches rrweb events (primary tab) for programmatic replay. Use --no-record to disable.\nLogging is ON by default. Use get-logs to retrieve logs; disable with --no-logs.\nConnection timeout: 5 minutes to connect after creation before auto-termination.\nKeep-alive sessions survive disconnections and must be explicitly terminated.\nContext persistence: When a session was created with a context using persist=true (default), wait a few seconds after termination before creating a new session with the same context.\nNamed contexts: Use --name with create-context to save friendly names (e.g. github, slack). Use the name anywhere a context ID is expected.\nWorkspace state: Workspaces are stored locally under ~/.browserbase/workspaces/.json (or BROWSERBASE_CONFIG_DIR/workspaces). They include the context id, active session id, and the last saved tab snapshot.\nOne context per site: Use separate contexts for different authenticated sites.\nAvoid concurrent sessions on the same context.\nRegions: us-west-2 (default), us-east-1, eu-central-1, ap-southeast-1.\nSession timeout: 60–21600 seconds (max 6 hours).\nCosts/limits: Your Browserbase plan has limits (browser hours, proxy data, concurrency). Keep-alive sessions consume hours while running; terminate sessions and set reasonable --timeout values to control cost. Check your Browserbase dashboard for current quotas.\nError Handling\n\nAll commands return JSON output. On error, the output includes an \"error\" key. Common errors:\n\nAPIConnectionError: Browserbase API unreachable\nRateLimitError: Too many concurrent sessions for your plan\nAPIStatusError: Invalid parameters or authentication failure\nMissing env vars: Set BROWSERBASE_API_KEY and BROWSERBASE_PROJECT_ID\nReference\n\nFor full API details, read {baseDir}/references/api-quick-ref.md." }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/JamesFincher/browserbase-sessions", "publisherUrl": "https://clawhub.ai/JamesFincher/browserbase-sessions", "owner": "JamesFincher", "version": "2.5.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/browserbase-sessions", "downloadUrl": "https://openagent3.xyz/downloads/browserbase-sessions", "agentUrl": "https://openagent3.xyz/skills/browserbase-sessions/agent", "manifestUrl": "https://openagent3.xyz/skills/browserbase-sessions/agent.json", "briefUrl": "https://openagent3.xyz/skills/browserbase-sessions/agent.md" } }