{
  "schemaVersion": "1.0",
  "item": {
    "slug": "e2b-desktop",
    "name": "E2B Desktop",
    "source": "tencent",
    "type": "skill",
    "category": "AI 智能",
    "sourceUrl": "https://clawhub.ai/EYHN/e2b-desktop",
    "canonicalUrl": "https://clawhub.ai/EYHN/e2b-desktop",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/e2b-desktop",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=e2b-desktop",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "SKILL.md",
      "scripts/_connect.py",
      "scripts/click.sh",
      "scripts/double_click.sh",
      "scripts/drag.sh",
      "scripts/get_cursor.sh"
    ],
    "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. 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. Summarize what changed and any follow-up checks I should run."
        }
      ]
    },
    "sourceHealth": {
      "source": "tencent",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-04-30T16:55:25.780Z",
      "expiresAt": "2026-05-07T16:55:25.780Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network",
        "contentDisposition": "attachment; filename=\"network-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/e2b-desktop"
    },
    "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/e2b-desktop",
    "agentPageUrl": "https://openagent3.xyz/skills/e2b-desktop/agent",
    "manifestUrl": "https://openagent3.xyz/skills/e2b-desktop/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/e2b-desktop/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. 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. Summarize what changed and any follow-up checks I should run."
      }
    ]
  },
  "documentation": {
    "source": "clawhub",
    "primaryDoc": "SKILL.md",
    "sections": [
      {
        "title": "E2B Desktop Skill",
        "body": "Control a headless Linux desktop (Ubuntu + XFCE) via the e2b-desktop Python SDK.\nAll scripts live in scripts/ and wrap the SDK in bash for easy agent use."
      },
      {
        "title": "Prerequisites",
        "body": "pip install e2b-desktop\nexport E2B_API_KEY=e2b_***"
      },
      {
        "title": "State Management",
        "body": "start_sandbox.sh saves the sandbox ID to ~/.e2b_state\nAll other scripts auto-load it from there\nOverride anytime with export E2B_SANDBOX_ID=<id>\nSandboxes survive script exit — reconnect with Sandbox.connect(sandbox_id)"
      },
      {
        "title": "Scripts",
        "body": "ScriptUsageDescriptionstart_sandbox.sh[--resolution 1280x800] [--timeout 300] [--stream]Create sandbox; optionally start VNC streamkill_sandbox.sh[SANDBOX_ID]Kill sandbox and remove statescreenshot.sh[OUTPUT_FILE]Take screenshot → PNG (default: /tmp/e2b_screenshot.png)click.shX YLeft click at coordinatesright_click.shX YRight clickdouble_click.shX YDouble clickmiddle_click.shX YMiddle clickmove_mouse.shX YMove cursor (no click)drag.shX1 Y1 X2 Y2Click-drag between two pointsscroll.shAMOUNTScroll (positive=up, negative=down)type_text.sh\"text\"Type text at current cursorpress_key.shKEY [KEY2...]Press key or combo (e.g. ctrl c)run_command.sh\"cmd\"Run shell command inside sandboxopen_url.shURL_OR_PATHOpen URL or file in default applaunch_app.shAPP_NAMELaunch app (e.g. firefox, vscode)stream_start.sh[--auth]Start VNC stream; --auth for password-protectedstream_stop.sh(none)Stop VNC streamget_cursor.sh(none)Print CURSOR_X and CURSOR_Yget_screen_size.sh(none)Print SCREEN_WIDTH and SCREEN_HEIGHTlist_windows.sh[APP_NAME]List app windows or show active windowwait.shMILLISECONDSWait N ms (sandbox-side)"
      },
      {
        "title": "Computer-Use Agent Loop Pattern",
        "body": "SCRIPTS=\"skills/e2b-desktop/scripts\"\n\n# 1. Start sandbox\nsource <($SCRIPTS/start_sandbox.sh --resolution 1280x800 --stream)\necho \"Sandbox: $SANDBOX_ID\"\necho \"View at: $STREAM_URL\"\n\n# 2. Agent loop\nwhile true; do\n  # Capture screen\n  $SCRIPTS/screenshot.sh /tmp/screen.png\n\n  # Send to LLM, parse action... (your code)\n  ACTION=$(llm_decide /tmp/screen.png)\n\n  case \"$ACTION\" in\n    click:*)   IFS=: read -r _ x y <<< \"$ACTION\"; $SCRIPTS/click.sh $x $y ;;\n    type:*)    $SCRIPTS/type_text.sh \"${ACTION#type:}\" ;;\n    key:*)     $SCRIPTS/press_key.sh ${ACTION#key:} ;;\n    done)      break ;;\n  esac\ndone\n\n# 3. Clean up\n$SCRIPTS/kill_sandbox.sh"
      },
      {
        "title": "Key Notes",
        "body": "scroll.sh AMOUNT: positive = scroll up, negative = scroll down (matches desktop.scroll(amount) API)\npress_key.sh ctrl c: multiple args become a key combo via desktop.press([\"ctrl\", \"c\"])\nrun_command.sh exits with the sandbox command's exit code\nAll mouse coordinate scripts accept integer pixel coordinates matching sandbox resolution\nVNC stream: only one active stream at a time; stop before switching windows"
      }
    ],
    "body": "E2B Desktop Skill\n\nControl a headless Linux desktop (Ubuntu + XFCE) via the e2b-desktop Python SDK. All scripts live in scripts/ and wrap the SDK in bash for easy agent use.\n\nPrerequisites\npip install e2b-desktop\nexport E2B_API_KEY=e2b_***\n\nState Management\nstart_sandbox.sh saves the sandbox ID to ~/.e2b_state\nAll other scripts auto-load it from there\nOverride anytime with export E2B_SANDBOX_ID=<id>\nSandboxes survive script exit — reconnect with Sandbox.connect(sandbox_id)\nScripts\nScript\tUsage\tDescription\nstart_sandbox.sh\t[--resolution 1280x800] [--timeout 300] [--stream]\tCreate sandbox; optionally start VNC stream\nkill_sandbox.sh\t[SANDBOX_ID]\tKill sandbox and remove state\nscreenshot.sh\t[OUTPUT_FILE]\tTake screenshot → PNG (default: /tmp/e2b_screenshot.png)\nclick.sh\tX Y\tLeft click at coordinates\nright_click.sh\tX Y\tRight click\ndouble_click.sh\tX Y\tDouble click\nmiddle_click.sh\tX Y\tMiddle click\nmove_mouse.sh\tX Y\tMove cursor (no click)\ndrag.sh\tX1 Y1 X2 Y2\tClick-drag between two points\nscroll.sh\tAMOUNT\tScroll (positive=up, negative=down)\ntype_text.sh\t\"text\"\tType text at current cursor\npress_key.sh\tKEY [KEY2...]\tPress key or combo (e.g. ctrl c)\nrun_command.sh\t\"cmd\"\tRun shell command inside sandbox\nopen_url.sh\tURL_OR_PATH\tOpen URL or file in default app\nlaunch_app.sh\tAPP_NAME\tLaunch app (e.g. firefox, vscode)\nstream_start.sh\t[--auth]\tStart VNC stream; --auth for password-protected\nstream_stop.sh\t(none)\tStop VNC stream\nget_cursor.sh\t(none)\tPrint CURSOR_X and CURSOR_Y\nget_screen_size.sh\t(none)\tPrint SCREEN_WIDTH and SCREEN_HEIGHT\nlist_windows.sh\t[APP_NAME]\tList app windows or show active window\nwait.sh\tMILLISECONDS\tWait N ms (sandbox-side)\nComputer-Use Agent Loop Pattern\nSCRIPTS=\"skills/e2b-desktop/scripts\"\n\n# 1. Start sandbox\nsource <($SCRIPTS/start_sandbox.sh --resolution 1280x800 --stream)\necho \"Sandbox: $SANDBOX_ID\"\necho \"View at: $STREAM_URL\"\n\n# 2. Agent loop\nwhile true; do\n  # Capture screen\n  $SCRIPTS/screenshot.sh /tmp/screen.png\n\n  # Send to LLM, parse action... (your code)\n  ACTION=$(llm_decide /tmp/screen.png)\n\n  case \"$ACTION\" in\n    click:*)   IFS=: read -r _ x y <<< \"$ACTION\"; $SCRIPTS/click.sh $x $y ;;\n    type:*)    $SCRIPTS/type_text.sh \"${ACTION#type:}\" ;;\n    key:*)     $SCRIPTS/press_key.sh ${ACTION#key:} ;;\n    done)      break ;;\n  esac\ndone\n\n# 3. Clean up\n$SCRIPTS/kill_sandbox.sh\n\nKey Notes\nscroll.sh AMOUNT: positive = scroll up, negative = scroll down (matches desktop.scroll(amount) API)\npress_key.sh ctrl c: multiple args become a key combo via desktop.press([\"ctrl\", \"c\"])\nrun_command.sh exits with the sandbox command's exit code\nAll mouse coordinate scripts accept integer pixel coordinates matching sandbox resolution\nVNC stream: only one active stream at a time; stop before switching windows"
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/EYHN/e2b-desktop",
    "publisherUrl": "https://clawhub.ai/EYHN/e2b-desktop",
    "owner": "EYHN",
    "version": "1.0.0",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/e2b-desktop",
    "downloadUrl": "https://openagent3.xyz/downloads/e2b-desktop",
    "agentUrl": "https://openagent3.xyz/skills/e2b-desktop/agent",
    "manifestUrl": "https://openagent3.xyz/skills/e2b-desktop/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/e2b-desktop/agent.md"
  }
}