{
  "schemaVersion": "1.0",
  "item": {
    "slug": "claude-team",
    "name": "Claude Team",
    "source": "tencent",
    "type": "skill",
    "category": "开发工具",
    "sourceUrl": "https://clawhub.ai/jalehman/claude-team",
    "canonicalUrl": "https://clawhub.ai/jalehman/claude-team",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/claude-team",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=claude-team",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "SKILL.md",
      "assets/setup.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/claude-team"
    },
    "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/claude-team",
    "agentPageUrl": "https://openagent3.xyz/skills/claude-team/agent",
    "manifestUrl": "https://openagent3.xyz/skills/claude-team/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/claude-team/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": "Claude Team",
        "body": "Claude-team is an MCP server that lets you spawn and manage teams of Claude Code sessions via iTerm2. Each worker gets their own terminal pane, optional git worktree, and can be assigned beads issues."
      },
      {
        "title": "Why Use Claude Team?",
        "body": "Parallelism: Fan out work to multiple agents working simultaneously\nContext isolation: Each worker has fresh context, keeps coordinator context clean\nVisibility: Real Claude Code sessions you can watch, interrupt, or take over\nGit worktrees: Each worker can have an isolated branch for their work"
      },
      {
        "title": "⚠️ Important Rule",
        "body": "NEVER make code changes directly. Always spawn workers for code changes. This keeps your context clean and provides proper git workflow with worktrees."
      },
      {
        "title": "Prerequisites",
        "body": "macOS with iTerm2 (Python API enabled: Preferences → General → Magic → Enable Python API)\nclaude-team MCP server configured in ~/.claude.json"
      },
      {
        "title": "Using via mcporter",
        "body": "All tools are called through mcporter call claude-team.<tool>:\n\nmcporter call claude-team.list_workers\nmcporter call claude-team.spawn_workers workers='[{\"project_path\":\"/path/to/repo\",\"bead\":\"cp-123\"}]'"
      },
      {
        "title": "spawn_workers",
        "body": "Create new Claude Code worker sessions.\n\nmcporter call claude-team.spawn_workers \\\n  workers='[{\n    \"project_path\": \"/path/to/repo\",\n    \"bead\": \"cp-123\",\n    \"annotation\": \"Fix auth bug\",\n    \"use_worktree\": true,\n    \"skip_permissions\": true\n  }]' \\\n  layout=\"auto\"\n\nWorker config fields:\n\nproject_path: Required. Path to repo or \"auto\" (uses CLAUDE_TEAM_PROJECT_DIR)\nbead: Optional beads issue ID — worker will follow beads workflow\nannotation: Task description (shown on badge, used in branch name)\nprompt: Additional instructions (if no bead, this is their assignment)\nuse_worktree: Create isolated git worktree (default: true)\nskip_permissions: Start with --dangerously-skip-permissions (default: false)\nname: Optional worker name override (auto-picks from themed sets otherwise)\n\nLayout options:\n\n\"auto\": Reuse existing claude-team windows, split into available space\n\"new\": Always create fresh window (1-4 workers in grid layout)"
      },
      {
        "title": "list_workers",
        "body": "See all managed workers:\n\nmcporter call claude-team.list_workers\nmcporter call claude-team.list_workers status_filter=\"ready\"\n\nStatus values: spawning, ready, busy, closed"
      },
      {
        "title": "message_workers",
        "body": "Send messages to one or more workers:\n\nmcporter call claude-team.message_workers \\\n  session_ids='[\"Groucho\"]' \\\n  message=\"Please also add unit tests\" \\\n  wait_mode=\"none\"\n\nwait_mode options:\n\n\"none\": Fire and forget (default)\n\"any\": Return when any worker is idle\n\"all\": Return when all workers are idle"
      },
      {
        "title": "check_idle_workers / wait_idle_workers",
        "body": "Check or wait for workers to finish:\n\n# Quick poll\nmcporter call claude-team.check_idle_workers session_ids='[\"Groucho\",\"Harpo\"]'\n\n# Blocking wait\nmcporter call claude-team.wait_idle_workers \\\n  session_ids='[\"Groucho\",\"Harpo\"]' \\\n  mode=\"all\" \\\n  timeout=600"
      },
      {
        "title": "read_worker_logs",
        "body": "Get conversation history:\n\nmcporter call claude-team.read_worker_logs \\\n  session_id=\"Groucho\" \\\n  pages=2"
      },
      {
        "title": "examine_worker",
        "body": "Get detailed status including conversation stats:\n\nmcporter call claude-team.examine_worker session_id=\"Groucho\""
      },
      {
        "title": "close_workers",
        "body": "Terminate workers when done:\n\nmcporter call claude-team.close_workers session_ids='[\"Groucho\",\"Harpo\"]'\n\n⚠️ Worktree cleanup: Workers with worktrees commit to ephemeral branches. After closing:\n\nReview commits on the worker's branch\nMerge or cherry-pick to a persistent branch\nDelete the branch: git branch -D <branch-name>"
      },
      {
        "title": "bd_help",
        "body": "Quick reference for beads commands:\n\nmcporter call claude-team.bd_help"
      },
      {
        "title": "Worker Identification",
        "body": "Workers can be referenced by any of:\n\nInternal ID: Short hex string (e.g., 3962c5c4)\nTerminal ID: iterm:UUID format\nWorker name: Human-friendly name (e.g., Groucho, Aragorn)"
      },
      {
        "title": "Workflow: Assigning a Beads Issue",
        "body": "# 1. Spawn worker with a bead assignment\nmcporter call claude-team.spawn_workers \\\n  workers='[{\n    \"project_path\": \"/Users/phaedrus/Projects/myrepo\",\n    \"bead\": \"proj-abc\",\n    \"annotation\": \"Implement config schemas\",\n    \"use_worktree\": true,\n    \"skip_permissions\": true\n  }]'\n\n# 2. Worker automatically:\n#    - Creates worktree with branch named after bead\n#    - Runs `bd show proj-abc` to understand the task\n#    - Marks issue in_progress\n#    - Implements the work\n#    - Closes the issue\n#    - Commits with issue reference\n\n# 3. Monitor progress\nmcporter call claude-team.check_idle_workers session_ids='[\"Groucho\"]'\nmcporter call claude-team.read_worker_logs session_id=\"Groucho\"\n\n# 4. When done, close and merge\nmcporter call claude-team.close_workers session_ids='[\"Groucho\"]'\n# Then: git merge or cherry-pick from worker's branch"
      },
      {
        "title": "Workflow: Parallel Fan-Out",
        "body": "# Spawn multiple workers for parallel tasks\nmcporter call claude-team.spawn_workers \\\n  workers='[\n    {\"project_path\": \"auto\", \"bead\": \"cp-123\", \"annotation\": \"Auth module\"},\n    {\"project_path\": \"auto\", \"bead\": \"cp-124\", \"annotation\": \"API routes\"},\n    {\"project_path\": \"auto\", \"bead\": \"cp-125\", \"annotation\": \"Unit tests\"}\n  ]' \\\n  layout=\"new\"\n\n# Wait for all to complete\nmcporter call claude-team.wait_idle_workers \\\n  session_ids='[\"Groucho\",\"Harpo\",\"Chico\"]' \\\n  mode=\"all\"\n\n# Review and close\nmcporter call claude-team.close_workers \\\n  session_ids='[\"Groucho\",\"Harpo\",\"Chico\"]'"
      },
      {
        "title": "Best Practices",
        "body": "Use beads: Assign bead IDs so workers follow proper issue workflow\nUse worktrees: Keeps work isolated, enables parallel commits\nSkip permissions: Workers need skip_permissions: true to write files\nMonitor, don't micromanage: Let workers complete, then review\nMerge carefully: Review worker branches before merging to main\nClose workers: Always close when done to clean up worktrees"
      },
      {
        "title": "HTTP Mode (Streamable HTTP Transport)",
        "body": "For persistent server operation, claude-team can run as an HTTP server. This keeps the MCP server running continuously with persistent state, avoiding cold starts."
      },
      {
        "title": "Starting the HTTP Server",
        "body": "Run the claude-team HTTP server directly:\n\n# From the claude-team directory\nuv run python -m claude_team_mcp --http --port 8766\n\n# Or specify the directory explicitly\nuv run --directory /path/to/claude-team python -m claude_team_mcp --http --port 8766\n\nFor automatic startup on login, use launchd (see the \"launchd Auto-Start\" section below)."
      },
      {
        "title": "mcporter.json Configuration",
        "body": "Once the HTTP server is running, configure mcporter to connect to it. Create ~/.mcporter/mcporter.json:\n\n{\n  \"mcpServers\": {\n    \"claude-team\": {\n      \"transport\": \"streamable-http\",\n      \"url\": \"http://127.0.0.1:8766/mcp\",\n      \"lifecycle\": \"keep-alive\"\n    }\n  }\n}"
      },
      {
        "title": "Benefits of HTTP Mode",
        "body": "Persistent state: Worker registry survives across CLI invocations\nFaster responses: No Python environment startup on each call\nExternal access: Can be accessed by cron jobs, scripts, or other tools\nSession recovery: Server tracks sessions even if coordinator disconnects"
      },
      {
        "title": "Connecting from Claude Code",
        "body": "Update your .mcp.json to use HTTP transport:\n\n{\n  \"mcpServers\": {\n    \"claude-team\": {\n      \"transport\": \"streamable-http\",\n      \"url\": \"http://127.0.0.1:8766/mcp\"\n    }\n  }\n}"
      },
      {
        "title": "launchd Auto-Start",
        "body": "To automatically start the claude-team server on login, use the bundled setup script."
      },
      {
        "title": "Quick Setup",
        "body": "Run the setup script from the skill's assets directory:\n\n# From the skill directory\n./assets/setup.sh\n\n# Or specify a custom claude-team location\nCLAUDE_TEAM_DIR=/path/to/claude-team ./assets/setup.sh"
      },
      {
        "title": "What the Setup Does",
        "body": "The setup script:\n\nDetects your uv installation path\nCreates the log directory at ~/.claude-team/logs/\nGenerates a launchd plist from assets/com.claude-team.plist.template\nInstalls it to ~/Library/LaunchAgents/com.claude-team.plist\nLoads the service to start immediately\n\nThe plist template uses uv run to start the HTTP server on port 8766, configured for iTerm2 Python API access (Aqua session type)."
      },
      {
        "title": "Managing the Service",
        "body": "# Stop the service\nlaunchctl unload ~/Library/LaunchAgents/com.claude-team.plist\n\n# Restart (re-run setup)\n./assets/setup.sh\n\n# Check if running\nlaunchctl list | grep claude-team\n\n# View logs\ntail -f ~/.claude-team/logs/stdout.log\ntail -f ~/.claude-team/logs/stderr.log"
      },
      {
        "title": "Troubleshooting launchd",
        "body": "# Check for load errors\nlaunchctl print gui/$UID/com.claude-team\n\n# Force restart\nlaunchctl kickstart -k gui/$UID/com.claude-team\n\n# Remove and reload (if plist changed)\nlaunchctl bootout gui/$UID/com.claude-team\nlaunchctl bootstrap gui/$UID ~/Library/LaunchAgents/com.claude-team.plist"
      },
      {
        "title": "Cron Integration",
        "body": "For background monitoring and notifications, claude-team supports cron-based worker tracking."
      },
      {
        "title": "Worker Tracking File",
        "body": "Claude-team writes worker state to ~/.claude-team/memory/worker-tracking.json:\n\n{\n  \"workers\": {\n    \"Groucho\": {\n      \"session_id\": \"3962c5c4\",\n      \"bead\": \"cp-123\",\n      \"annotation\": \"Fix auth bug\",\n      \"status\": \"busy\",\n      \"project_path\": \"/Users/phaedrus/Projects/myrepo\",\n      \"started_at\": \"2025-01-05T10:30:00Z\",\n      \"last_activity\": \"2025-01-05T11:45:00Z\"\n    },\n    \"Harpo\": {\n      \"session_id\": \"a1b2c3d4\",\n      \"bead\": \"cp-124\",\n      \"annotation\": \"Add API routes\",\n      \"status\": \"idle\",\n      \"project_path\": \"/Users/phaedrus/Projects/myrepo\",\n      \"started_at\": \"2025-01-05T10:30:00Z\",\n      \"last_activity\": \"2025-01-05T11:50:00Z\",\n      \"completed_at\": \"2025-01-05T11:50:00Z\"\n    }\n  },\n  \"last_updated\": \"2025-01-05T11:50:00Z\"\n}"
      },
      {
        "title": "Cron Job for Monitoring Completions",
        "body": "Create a monitoring script at ~/.claude-team/scripts/check-workers.sh:\n\n#!/bin/bash\n# Check for completed workers and send notifications\n\nTRACKING_FILE=\"$HOME/.claude-team/memory/worker-tracking.json\"\nNOTIFIED_FILE=\"$HOME/.claude-team/memory/notified-workers.json\"\nTELEGRAM_BOT_TOKEN=\"${TELEGRAM_BOT_TOKEN}\"\nTELEGRAM_CHAT_ID=\"${TELEGRAM_CHAT_ID}\"\n\n# Exit if tracking file doesn't exist\n[ -f \"$TRACKING_FILE\" ] || exit 0\n\n# Initialize notified file if needed\n[ -f \"$NOTIFIED_FILE\" ] || echo '{\"notified\":[]}' > \"$NOTIFIED_FILE\"\n\n# Find idle workers that haven't been notified\nIDLE_WORKERS=$(jq -r '\n  .workers | to_entries[] |\n  select(.value.status == \"idle\") |\n  .key\n' \"$TRACKING_FILE\")\n\nfor worker in $IDLE_WORKERS; do\n  # Check if already notified\n  ALREADY_NOTIFIED=$(jq -r --arg w \"$worker\" '.notified | index($w) != null' \"$NOTIFIED_FILE\")\n\n  if [ \"$ALREADY_NOTIFIED\" = \"false\" ]; then\n    # Get worker details\n    BEAD=$(jq -r --arg w \"$worker\" '.workers[$w].bead // \"no-bead\"' \"$TRACKING_FILE\")\n    ANNOTATION=$(jq -r --arg w \"$worker\" '.workers[$w].annotation // \"no annotation\"' \"$TRACKING_FILE\")\n\n    # Send Telegram notification\n    MESSAGE=\"🤖 Worker *${worker}* completed\n📋 Bead: \\`${BEAD}\\`\n📝 ${ANNOTATION}\"\n\n    curl -s -X POST \"https://api.telegram.org/bot${TELEGRAM_BOT_TOKEN}/sendMessage\" \\\n      -d chat_id=\"$TELEGRAM_CHAT_ID\" \\\n      -d text=\"$MESSAGE\" \\\n      -d parse_mode=\"Markdown\" > /dev/null\n\n    # Mark as notified\n    jq --arg w \"$worker\" '.notified += [$w]' \"$NOTIFIED_FILE\" > \"${NOTIFIED_FILE}.tmp\"\n    mv \"${NOTIFIED_FILE}.tmp\" \"$NOTIFIED_FILE\"\n  fi\ndone\n\nMake it executable:\n\nchmod +x ~/.claude-team/scripts/check-workers.sh"
      },
      {
        "title": "Crontab Entry",
        "body": "Add to crontab (crontab -e):\n\n# Check claude-team workers every 2 minutes\n*/2 * * * * TELEGRAM_BOT_TOKEN=\"your-bot-token\" TELEGRAM_CHAT_ID=\"your-chat-id\" ~/.claude-team/scripts/check-workers.sh"
      },
      {
        "title": "Environment Setup",
        "body": "Set Telegram credentials in your shell profile (~/.zshrc):\n\nexport TELEGRAM_BOT_TOKEN=\"123456789:ABCdefGHIjklMNOpqrsTUVwxyz\"\nexport TELEGRAM_CHAT_ID=\"-1001234567890\""
      },
      {
        "title": "Alternative: Using clawdbot for Notifications",
        "body": "If you have clawdbot configured, you can send notifications through it instead:\n\n# In check-workers.sh, replace the curl command with:\nclawdbot send --to \"$TELEGRAM_CHAT_ID\" --message \"$MESSAGE\" --provider telegram"
      },
      {
        "title": "Clearing Notification State",
        "body": "When starting a fresh batch of workers, clear the notified list:\n\necho '{\"notified\":[]}' > ~/.claude-team/memory/notified-workers.json"
      }
    ],
    "body": "Claude Team\n\nClaude-team is an MCP server that lets you spawn and manage teams of Claude Code sessions via iTerm2. Each worker gets their own terminal pane, optional git worktree, and can be assigned beads issues.\n\nWhy Use Claude Team?\nParallelism: Fan out work to multiple agents working simultaneously\nContext isolation: Each worker has fresh context, keeps coordinator context clean\nVisibility: Real Claude Code sessions you can watch, interrupt, or take over\nGit worktrees: Each worker can have an isolated branch for their work\n⚠️ Important Rule\n\nNEVER make code changes directly. Always spawn workers for code changes. This keeps your context clean and provides proper git workflow with worktrees.\n\nPrerequisites\nmacOS with iTerm2 (Python API enabled: Preferences → General → Magic → Enable Python API)\nclaude-team MCP server configured in ~/.claude.json\nUsing via mcporter\n\nAll tools are called through mcporter call claude-team.<tool>:\n\nmcporter call claude-team.list_workers\nmcporter call claude-team.spawn_workers workers='[{\"project_path\":\"/path/to/repo\",\"bead\":\"cp-123\"}]'\n\nCore Tools\nspawn_workers\n\nCreate new Claude Code worker sessions.\n\nmcporter call claude-team.spawn_workers \\\n  workers='[{\n    \"project_path\": \"/path/to/repo\",\n    \"bead\": \"cp-123\",\n    \"annotation\": \"Fix auth bug\",\n    \"use_worktree\": true,\n    \"skip_permissions\": true\n  }]' \\\n  layout=\"auto\"\n\n\nWorker config fields:\n\nproject_path: Required. Path to repo or \"auto\" (uses CLAUDE_TEAM_PROJECT_DIR)\nbead: Optional beads issue ID — worker will follow beads workflow\nannotation: Task description (shown on badge, used in branch name)\nprompt: Additional instructions (if no bead, this is their assignment)\nuse_worktree: Create isolated git worktree (default: true)\nskip_permissions: Start with --dangerously-skip-permissions (default: false)\nname: Optional worker name override (auto-picks from themed sets otherwise)\n\nLayout options:\n\n\"auto\": Reuse existing claude-team windows, split into available space\n\"new\": Always create fresh window (1-4 workers in grid layout)\nlist_workers\n\nSee all managed workers:\n\nmcporter call claude-team.list_workers\nmcporter call claude-team.list_workers status_filter=\"ready\"\n\n\nStatus values: spawning, ready, busy, closed\n\nmessage_workers\n\nSend messages to one or more workers:\n\nmcporter call claude-team.message_workers \\\n  session_ids='[\"Groucho\"]' \\\n  message=\"Please also add unit tests\" \\\n  wait_mode=\"none\"\n\n\nwait_mode options:\n\n\"none\": Fire and forget (default)\n\"any\": Return when any worker is idle\n\"all\": Return when all workers are idle\ncheck_idle_workers / wait_idle_workers\n\nCheck or wait for workers to finish:\n\n# Quick poll\nmcporter call claude-team.check_idle_workers session_ids='[\"Groucho\",\"Harpo\"]'\n\n# Blocking wait\nmcporter call claude-team.wait_idle_workers \\\n  session_ids='[\"Groucho\",\"Harpo\"]' \\\n  mode=\"all\" \\\n  timeout=600\n\nread_worker_logs\n\nGet conversation history:\n\nmcporter call claude-team.read_worker_logs \\\n  session_id=\"Groucho\" \\\n  pages=2\n\nexamine_worker\n\nGet detailed status including conversation stats:\n\nmcporter call claude-team.examine_worker session_id=\"Groucho\"\n\nclose_workers\n\nTerminate workers when done:\n\nmcporter call claude-team.close_workers session_ids='[\"Groucho\",\"Harpo\"]'\n\n\n⚠️ Worktree cleanup: Workers with worktrees commit to ephemeral branches. After closing:\n\nReview commits on the worker's branch\nMerge or cherry-pick to a persistent branch\nDelete the branch: git branch -D <branch-name>\nbd_help\n\nQuick reference for beads commands:\n\nmcporter call claude-team.bd_help\n\nWorker Identification\n\nWorkers can be referenced by any of:\n\nInternal ID: Short hex string (e.g., 3962c5c4)\nTerminal ID: iterm:UUID format\nWorker name: Human-friendly name (e.g., Groucho, Aragorn)\nWorkflow: Assigning a Beads Issue\n# 1. Spawn worker with a bead assignment\nmcporter call claude-team.spawn_workers \\\n  workers='[{\n    \"project_path\": \"/Users/phaedrus/Projects/myrepo\",\n    \"bead\": \"proj-abc\",\n    \"annotation\": \"Implement config schemas\",\n    \"use_worktree\": true,\n    \"skip_permissions\": true\n  }]'\n\n# 2. Worker automatically:\n#    - Creates worktree with branch named after bead\n#    - Runs `bd show proj-abc` to understand the task\n#    - Marks issue in_progress\n#    - Implements the work\n#    - Closes the issue\n#    - Commits with issue reference\n\n# 3. Monitor progress\nmcporter call claude-team.check_idle_workers session_ids='[\"Groucho\"]'\nmcporter call claude-team.read_worker_logs session_id=\"Groucho\"\n\n# 4. When done, close and merge\nmcporter call claude-team.close_workers session_ids='[\"Groucho\"]'\n# Then: git merge or cherry-pick from worker's branch\n\nWorkflow: Parallel Fan-Out\n# Spawn multiple workers for parallel tasks\nmcporter call claude-team.spawn_workers \\\n  workers='[\n    {\"project_path\": \"auto\", \"bead\": \"cp-123\", \"annotation\": \"Auth module\"},\n    {\"project_path\": \"auto\", \"bead\": \"cp-124\", \"annotation\": \"API routes\"},\n    {\"project_path\": \"auto\", \"bead\": \"cp-125\", \"annotation\": \"Unit tests\"}\n  ]' \\\n  layout=\"new\"\n\n# Wait for all to complete\nmcporter call claude-team.wait_idle_workers \\\n  session_ids='[\"Groucho\",\"Harpo\",\"Chico\"]' \\\n  mode=\"all\"\n\n# Review and close\nmcporter call claude-team.close_workers \\\n  session_ids='[\"Groucho\",\"Harpo\",\"Chico\"]'\n\nBest Practices\nUse beads: Assign bead IDs so workers follow proper issue workflow\nUse worktrees: Keeps work isolated, enables parallel commits\nSkip permissions: Workers need skip_permissions: true to write files\nMonitor, don't micromanage: Let workers complete, then review\nMerge carefully: Review worker branches before merging to main\nClose workers: Always close when done to clean up worktrees\nHTTP Mode (Streamable HTTP Transport)\n\nFor persistent server operation, claude-team can run as an HTTP server. This keeps the MCP server running continuously with persistent state, avoiding cold starts.\n\nStarting the HTTP Server\n\nRun the claude-team HTTP server directly:\n\n# From the claude-team directory\nuv run python -m claude_team_mcp --http --port 8766\n\n# Or specify the directory explicitly\nuv run --directory /path/to/claude-team python -m claude_team_mcp --http --port 8766\n\n\nFor automatic startup on login, use launchd (see the \"launchd Auto-Start\" section below).\n\nmcporter.json Configuration\n\nOnce the HTTP server is running, configure mcporter to connect to it. Create ~/.mcporter/mcporter.json:\n\n{\n  \"mcpServers\": {\n    \"claude-team\": {\n      \"transport\": \"streamable-http\",\n      \"url\": \"http://127.0.0.1:8766/mcp\",\n      \"lifecycle\": \"keep-alive\"\n    }\n  }\n}\n\nBenefits of HTTP Mode\nPersistent state: Worker registry survives across CLI invocations\nFaster responses: No Python environment startup on each call\nExternal access: Can be accessed by cron jobs, scripts, or other tools\nSession recovery: Server tracks sessions even if coordinator disconnects\nConnecting from Claude Code\n\nUpdate your .mcp.json to use HTTP transport:\n\n{\n  \"mcpServers\": {\n    \"claude-team\": {\n      \"transport\": \"streamable-http\",\n      \"url\": \"http://127.0.0.1:8766/mcp\"\n    }\n  }\n}\n\nlaunchd Auto-Start\n\nTo automatically start the claude-team server on login, use the bundled setup script.\n\nQuick Setup\n\nRun the setup script from the skill's assets directory:\n\n# From the skill directory\n./assets/setup.sh\n\n# Or specify a custom claude-team location\nCLAUDE_TEAM_DIR=/path/to/claude-team ./assets/setup.sh\n\nWhat the Setup Does\n\nThe setup script:\n\nDetects your uv installation path\nCreates the log directory at ~/.claude-team/logs/\nGenerates a launchd plist from assets/com.claude-team.plist.template\nInstalls it to ~/Library/LaunchAgents/com.claude-team.plist\nLoads the service to start immediately\n\nThe plist template uses uv run to start the HTTP server on port 8766, configured for iTerm2 Python API access (Aqua session type).\n\nManaging the Service\n# Stop the service\nlaunchctl unload ~/Library/LaunchAgents/com.claude-team.plist\n\n# Restart (re-run setup)\n./assets/setup.sh\n\n# Check if running\nlaunchctl list | grep claude-team\n\n# View logs\ntail -f ~/.claude-team/logs/stdout.log\ntail -f ~/.claude-team/logs/stderr.log\n\nTroubleshooting launchd\n# Check for load errors\nlaunchctl print gui/$UID/com.claude-team\n\n# Force restart\nlaunchctl kickstart -k gui/$UID/com.claude-team\n\n# Remove and reload (if plist changed)\nlaunchctl bootout gui/$UID/com.claude-team\nlaunchctl bootstrap gui/$UID ~/Library/LaunchAgents/com.claude-team.plist\n\nCron Integration\n\nFor background monitoring and notifications, claude-team supports cron-based worker tracking.\n\nWorker Tracking File\n\nClaude-team writes worker state to ~/.claude-team/memory/worker-tracking.json:\n\n{\n  \"workers\": {\n    \"Groucho\": {\n      \"session_id\": \"3962c5c4\",\n      \"bead\": \"cp-123\",\n      \"annotation\": \"Fix auth bug\",\n      \"status\": \"busy\",\n      \"project_path\": \"/Users/phaedrus/Projects/myrepo\",\n      \"started_at\": \"2025-01-05T10:30:00Z\",\n      \"last_activity\": \"2025-01-05T11:45:00Z\"\n    },\n    \"Harpo\": {\n      \"session_id\": \"a1b2c3d4\",\n      \"bead\": \"cp-124\",\n      \"annotation\": \"Add API routes\",\n      \"status\": \"idle\",\n      \"project_path\": \"/Users/phaedrus/Projects/myrepo\",\n      \"started_at\": \"2025-01-05T10:30:00Z\",\n      \"last_activity\": \"2025-01-05T11:50:00Z\",\n      \"completed_at\": \"2025-01-05T11:50:00Z\"\n    }\n  },\n  \"last_updated\": \"2025-01-05T11:50:00Z\"\n}\n\nCron Job for Monitoring Completions\n\nCreate a monitoring script at ~/.claude-team/scripts/check-workers.sh:\n\n#!/bin/bash\n# Check for completed workers and send notifications\n\nTRACKING_FILE=\"$HOME/.claude-team/memory/worker-tracking.json\"\nNOTIFIED_FILE=\"$HOME/.claude-team/memory/notified-workers.json\"\nTELEGRAM_BOT_TOKEN=\"${TELEGRAM_BOT_TOKEN}\"\nTELEGRAM_CHAT_ID=\"${TELEGRAM_CHAT_ID}\"\n\n# Exit if tracking file doesn't exist\n[ -f \"$TRACKING_FILE\" ] || exit 0\n\n# Initialize notified file if needed\n[ -f \"$NOTIFIED_FILE\" ] || echo '{\"notified\":[]}' > \"$NOTIFIED_FILE\"\n\n# Find idle workers that haven't been notified\nIDLE_WORKERS=$(jq -r '\n  .workers | to_entries[] |\n  select(.value.status == \"idle\") |\n  .key\n' \"$TRACKING_FILE\")\n\nfor worker in $IDLE_WORKERS; do\n  # Check if already notified\n  ALREADY_NOTIFIED=$(jq -r --arg w \"$worker\" '.notified | index($w) != null' \"$NOTIFIED_FILE\")\n\n  if [ \"$ALREADY_NOTIFIED\" = \"false\" ]; then\n    # Get worker details\n    BEAD=$(jq -r --arg w \"$worker\" '.workers[$w].bead // \"no-bead\"' \"$TRACKING_FILE\")\n    ANNOTATION=$(jq -r --arg w \"$worker\" '.workers[$w].annotation // \"no annotation\"' \"$TRACKING_FILE\")\n\n    # Send Telegram notification\n    MESSAGE=\"🤖 Worker *${worker}* completed\n📋 Bead: \\`${BEAD}\\`\n📝 ${ANNOTATION}\"\n\n    curl -s -X POST \"https://api.telegram.org/bot${TELEGRAM_BOT_TOKEN}/sendMessage\" \\\n      -d chat_id=\"$TELEGRAM_CHAT_ID\" \\\n      -d text=\"$MESSAGE\" \\\n      -d parse_mode=\"Markdown\" > /dev/null\n\n    # Mark as notified\n    jq --arg w \"$worker\" '.notified += [$w]' \"$NOTIFIED_FILE\" > \"${NOTIFIED_FILE}.tmp\"\n    mv \"${NOTIFIED_FILE}.tmp\" \"$NOTIFIED_FILE\"\n  fi\ndone\n\n\nMake it executable:\n\nchmod +x ~/.claude-team/scripts/check-workers.sh\n\nCrontab Entry\n\nAdd to crontab (crontab -e):\n\n# Check claude-team workers every 2 minutes\n*/2 * * * * TELEGRAM_BOT_TOKEN=\"your-bot-token\" TELEGRAM_CHAT_ID=\"your-chat-id\" ~/.claude-team/scripts/check-workers.sh\n\nEnvironment Setup\n\nSet Telegram credentials in your shell profile (~/.zshrc):\n\nexport TELEGRAM_BOT_TOKEN=\"123456789:ABCdefGHIjklMNOpqrsTUVwxyz\"\nexport TELEGRAM_CHAT_ID=\"-1001234567890\"\n\nAlternative: Using clawdbot for Notifications\n\nIf you have clawdbot configured, you can send notifications through it instead:\n\n# In check-workers.sh, replace the curl command with:\nclawdbot send --to \"$TELEGRAM_CHAT_ID\" --message \"$MESSAGE\" --provider telegram\n\nClearing Notification State\n\nWhen starting a fresh batch of workers, clear the notified list:\n\necho '{\"notified\":[]}' > ~/.claude-team/memory/notified-workers.json"
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/jalehman/claude-team",
    "publisherUrl": "https://clawhub.ai/jalehman/claude-team",
    "owner": "jalehman",
    "version": "1.5.0",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/claude-team",
    "downloadUrl": "https://openagent3.xyz/downloads/claude-team",
    "agentUrl": "https://openagent3.xyz/skills/claude-team/agent",
    "manifestUrl": "https://openagent3.xyz/skills/claude-team/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/claude-team/agent.md"
  }
}