[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7F2GK2PQ01FY6JSW844F2Z4":3,"skill-same:01M7F2GK2PQ01FY6JSW844F2Z4":46},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":30,"updatedAt":31,"install":32,"resources":45},"01M7F2GK2PQ01FY6JSW844F2Z4","clawteam",null,"Multi-agent swarm coordination via the ClawTeam CLI. Use when the user wants to create agent teams, spawn multiple agents to work in parallel, coordinate tasks with dependencies, broadcast messages between agents, monitor progress via kanban board, or launch pre-built team templates (hedge-fund, code-review, research-paper). ClawTeam uses git worktree isolation + tmux + filesystem-based messaging. Trigger phrases: team, swarm, multi-agent, clawteam, spawn agents, parallel agents, agent team.","lza6\u002FArmor-piercing-system-prompts","lza6","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F128880206?v=4","01M7EB3K7WPX5T28EJ2NVHE6EG","Chinese-tingfeng","01M7EB3K97E1KQ16QWKJMFNGX8",11,"MIT","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F05-技能与代理开发\u002Fopenclaw\u002FSKILL.md",[],[],"2026-11-16 10:05:34",false,"---\nname: clawteam\ndescription: \"Multi-agent swarm coordination via the ClawTeam CLI. Use when the user wants to create agent teams, spawn multiple agents to work in parallel, coordinate tasks with dependencies, broadcast messages between agents, monitor progress via kanban board, or launch pre-built team templates (hedge-fund, code-review, research-paper). ClawTeam uses git worktree isolation + tmux + filesystem-based messaging. Trigger phrases: team, swarm, multi-agent, clawteam, spawn agents, parallel agents, agent team.\"\n---\n\n# ClawTeam — Multi-Agent Swarm Coordination\n\n## Overview\n\nClawTeam is a CLI tool (`clawteam`) for orchestrating multiple AI agents as self-organizing swarms. It uses git worktree isolation, tmux windows, and filesystem-based messaging. OpenClaw is the default agent backend.\n\n**CLI binary**: `clawteam` (installed via pip, available in PATH)\n\n## Quick Start\n\n### One-Command Template Launch (Recommended)\n\n```bash\n# Launch a pre-built team from a template\nclawteam launch hedge-fund --team fund1\nclawteam launch code-review --team review1\nclawteam launch research-paper --team paper1\n```\n\n### Manual Team Setup\n\n```bash\n# 1. Create team with leader\nclawteam team spawn-team my-team -d \"Build a web app\" -n leader\n\n# 2. Create tasks with dependencies\nclawteam task create my-team \"Design API schema\" -o architect\n# Returns task ID, e.g., abc123\n\nclawteam task create my-team \"Implement auth\" -o backend --blocked-by abc123\nclawteam task create my-team \"Build frontend\" -o frontend --blocked-by abc123\nclawteam task create my-team \"Write tests\" -o tester\n\n# 3. Spawn agents (each gets its own tmux window + git worktree)\nclawteam spawn -t my-team -n architect --task \"Design the API schema for a web app\"\nclawteam spawn -t my-team -n backend --task \"Implement OAuth2 authentication\"\nclawteam spawn -t my-team -n frontend --task \"Build React dashboard\"\n\n# 4. Monitor\nclawteam board show my-team        # Kanban view\nclawteam board attach my-team      # Tmux tiled view (all agents side-by-side)\nclawteam board serve --port 8080   # Web dashboard\n```\n\n## Command Reference\n\n### Team Management\n\n| Command | Description |\n|---------|-------------|\n| `clawteam team spawn-team \u003Cname> -d \"\u003Cdesc>\" -n \u003Cleader>` | Create team |\n| `clawteam team discover` | List all teams |\n| `clawteam team status \u003Cteam>` | Show team members and info |\n| `clawteam team cleanup \u003Cteam> --force` | Delete team and all data |\n\n### Task Management\n\n| Command | Description |\n|---------|-------------|\n| `clawteam task create \u003Cteam> \"\u003Csubject>\" -o \u003Cowner> [-d \"\u003Cdesc>\"] [--blocked-by \u003Cid>]` | Create task |\n| `clawteam task list \u003Cteam> [--owner \u003Cname>]` | List tasks (filterable) |\n| `clawteam task update \u003Cteam> \u003Cid> --status \u003Cstatus>` | Update status |\n| `clawteam task get \u003Cteam> \u003Cid>` | Get single task |\n| `clawteam task stats \u003Cteam>` | Timing statistics |\n| `clawteam task wait \u003Cteam>` | Block until all tasks complete |\n\n**Task statuses**: `pending`, `in_progress`, `completed`, `blocked`\n\n**Dependency auto-resolution**: When a blocking task completes, dependent tasks automatically change from `blocked` to `pending`.\n\n**Task locking**: When a task moves to `in_progress`, it is locked by the calling agent. Other agents cannot claim it unless they use `--force`. Stale locks from dead agents are automatically released.\n\n### Agent Spawning\n\n**IMPORTANT**: Always use the default command (`openclaw`) — do NOT override to `claude` or other agents. The default handles permissions, prompt injection, and nesting detection correctly. If you specify `claude` as the command, agents will get stuck on interactive permission prompts.\n\n```bash\n# Default (RECOMMENDED): spawns openclaw tui in tmux with prompt\nclawteam spawn -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask description>\"\n\n# Explicit backend (still uses openclaw by default)\nclawteam spawn tmux -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask>\"\nclawteam spawn subprocess -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask>\"\n\n# With git worktree isolation\nclawteam spawn -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask>\" --workspace --repo \u002Fpath\u002Fto\u002Frepo\n```\n\nEach spawned agent gets:\n- Its own tmux window (visible via `board attach`)\n- Its own git worktree branch (`clawteam\u002F{team}\u002F{agent}`)\n- An auto-injected coordination prompt (how to use clawteam CLI)\n- Environment variables: `CLAWTEAM_AGENT_NAME`, `CLAWTEAM_TEAM_NAME`, etc.\n\n**Spawn safety features:**\n- Commands are pre-validated before launch — you get a clear error if the agent CLI is not installed\n- If a spawn fails, the registered team member and worktree are automatically rolled back\n- Claude Code and Codex workspace trust prompts are auto-confirmed in fresh worktrees\n\n### Messaging\n\n| Command | Description |\n|---------|-------------|\n| `clawteam inbox send \u003Cteam> \u003Cto> \"\u003Cmsg>\" --from \u003Csender>` | Point-to-point message |\n| `clawteam inbox broadcast \u003Cteam> \"\u003Cmsg>\" --from \u003Csender>` | Broadcast to all |\n| `clawteam inbox peek \u003Cteam> -a \u003Cagent>` | Peek without consuming |\n| `clawteam inbox receive \u003Cteam>` | Consume messages |\n| `clawteam inbox log \u003Cteam>` | View message history |\n\n### Monitoring\n\n| Command | Description |\n|---------|-------------|\n| `clawteam board show \u003Cteam>` | Kanban board (rich terminal) |\n| `clawteam board overview` | All teams overview |\n| `clawteam board live \u003Cteam>` | Live-refreshing board |\n| `clawteam board attach \u003Cteam>` | Tmux tiled view |\n| `clawteam board serve --port 8080` | Web dashboard |\n\n### Cost Tracking\n\n| Command | Description |\n|---------|-------------|\n| `clawteam cost report \u003Cteam> --input-tokens \u003CN> --output-tokens \u003CN> --cost-cents \u003CN>` | Report usage |\n| `clawteam cost show \u003Cteam>` | Show summary |\n| `clawteam cost budget \u003Cteam> \u003Cdollars>` | Set budget |\n\n### Templates\n\n| Command | Description |\n|---------|-------------|\n| `clawteam template list` | List available templates |\n| `clawteam template show \u003Cname>` | Show template details |\n| `clawteam launch \u003Ctemplate> [--team-name \u003Cname>] [--goal \"\u003Cgoal>\"]` | Launch from template |\n\n**Built-in templates**: `hedge-fund`, `code-review`, `research-paper`\n\n### Configuration\n\n```bash\nclawteam config show                           # Show all settings\nclawteam config set transport file             # Set transport backend\nclawteam config set skip_permissions true      # Auto-skip permission prompts\nclawteam config health                         # System health check\n```\n\n### Other Commands\n\n| Command | Description |\n|---------|-------------|\n| `clawteam lifecycle idle \u003Cteam> --agent \u003Cname>` | Report agent idle |\n| `clawteam session save \u003Cteam> --session-id \u003Cid>` | Save session for resume |\n| `clawteam plan submit \u003Cteam> \"\u003Cplan>\" --from \u003Cagent>` | Submit plan for approval (team-scoped storage) |\n| `clawteam workspace list \u003Cteam>` | List git worktrees |\n| `clawteam workspace merge \u003Cteam> --agent \u003Cname>` | Merge agent branch |\n\n## JSON Output\n\nAdd `--json` before any subcommand for machine-readable output:\n\n```bash\nclawteam --json task list my-team\nclawteam --json team status my-team\n```\n\n## Typical Workflow\n\n1. **User says**: \"Create a team to build a web app\"\n2. **You do**: `clawteam team spawn-team webapp -d \"Build web app\" -n leader`\n3. **Create tasks**: Use `clawteam task create` with `--blocked-by` for dependencies\n4. **Spawn agents**: Use `clawteam spawn` for each worker\n5. **Monitor**: Start a background polling loop immediately — do NOT wait for user to ask\n6. **Communicate**: Use `clawteam inbox broadcast` for team-wide updates\n7. **Deliver**: Proactively send final results to the user as soon as all tasks complete\n8. **Cleanup**: `clawteam cost show`, `clawteam task stats`, merge worktrees, then `clawteam team cleanup webapp --force`\n\n## Leader Orchestration Pattern\n\nWhen YOU are the leader agent, follow this pattern to autonomously manage a swarm:\n\n### Phase 1: Analyze & Plan\n```\n1. Understand the user's goal\n2. Break it into independent subtasks\n3. Identify dependencies between tasks (what must finish before what)\n4. Decide how many worker agents are needed\n```\n\n### Phase 2: Setup\n```bash\n# Create team\nclawteam team spawn-team \u003Cteam> -d \"\u003Cgoal description>\" -n leader\n\n# Create tasks with dependency chains\nclawteam task create \u003Cteam> \"Design API\" -o architect\n# Save the returned task ID (e.g., abc123)\nclawteam task create \u003Cteam> \"Build backend\" -o backend --blocked-by abc123\nclawteam task create \u003Cteam> \"Build frontend\" -o frontend --blocked-by abc123\nclawteam task create \u003Cteam> \"Integration tests\" -o tester --blocked-by \u003Cbackend-id>,\u003Cfrontend-id>\n```\n\n### Phase 3: Spawn Workers\n```bash\n# Each spawn launches an openclaw tui in its own tmux window\nclawteam spawn -t \u003Cteam> -n architect --task \"Design REST API schema for \u003Cgoal>\"\nclawteam spawn -t \u003Cteam> -n backend --task \"Implement backend based on API schema\"\nclawteam spawn -t \u003Cteam> -n frontend --task \"Build React frontend\"\nclawteam spawn -t \u003Cteam> -n tester --task \"Write and run integration tests\"\n```\n\n### Phase 4: Monitor Loop\n\n**IMPORTANT**: Start monitoring immediately after spawning — do NOT wait for the user to ask for status updates. Run the monitor loop in the background right away so you can:\n1. **Push mid-progress updates proactively** — when ~50% of tasks complete, send the user a brief status update (e.g. \"4\u002F7 agents done, 3 still working\"). Do NOT wait for them to ask.\n2. **Deliver final results immediately** when all tasks complete.\n\n```bash\n# Poll task status every 30-60 seconds\nwhile true; do\n  clawteam --json task list \u003Cteam> | python3 -c \"\nimport sys, json\ntasks = json.load(sys.stdin)\ndone = sum(1 for t in tasks if t['status'] == 'completed')\ntotal = len(tasks)\nprint(f'{done}\u002F{total} complete')\nif done == total: print('ALL DONE'); sys.exit(0)\n\"\n  # Check for messages from workers\n  clawteam inbox receive \u003Cteam>\n  # IMPORTANT: Send a mid-progress update to the user when roughly half the tasks are done\n  sleep 30\ndone\n```\n\n### Phase 5: Converge & Report\n\n**IMPORTANT**: Proactively deliver results to the user as soon as all tasks complete. Do NOT wait for the user to ask. Include the final output, a summary, and cost\u002Ftiming stats. ALWAYS merge worktrees and clean up.\n\n```bash\n# After all tasks complete — do ALL of these steps:\nclawteam board show \u003Cteam>           # Final status\nclawteam cost show \u003Cteam>            # Total cost — include in report to user\nclawteam task stats \u003Cteam>           # Timing stats — include in report to user\n# Merge each worker's branch back to main\nfor agent in \u003Cagent1> \u003Cagent2> ...; do\n  clawteam workspace merge \u003Cteam> --agent $agent\ndone\nclawteam team cleanup \u003Cteam> --force  # Clean up — ALWAYS do this last\n# Then: send the final deliverables to the user immediately\n```\n\n### Decision Rules for the Leader\n- **Independent tasks** → spawn workers in parallel\n- **Sequential tasks** → use `--blocked-by` to chain them; ClawTeam auto-unblocks\n- **Worker asks for help** → check inbox, provide guidance via `inbox send`\n- **Worker stuck** → check task status; if `in_progress` too long, send a nudge via `inbox send`\n- **Worker done** → verify result via inbox message, then move to next phase\n- **All done** → merge worktrees, deliver results to user proactively, then cleanup\n- **Always** → start background monitoring immediately after spawn; never wait for user to ask for status\n\n## Data Location\n\nAll state stored in `~\u002F.clawteam\u002F`:\n- Teams: `~\u002F.clawteam\u002Fteams\u002F\u003Cteam>\u002Fconfig.json`\n- Tasks: `~\u002F.clawteam\u002Ftasks\u002F\u003Cteam>\u002Ftask-\u003Cid>.json` (with `fcntl` file locking for concurrent safety)\n- Plans: `~\u002F.clawteam\u002Fplans\u002F\u003Cteam>\u002F\u003Cagent>-\u003Cplan_id>.md` (team-scoped, isolated per team)\n- Messages: `~\u002F.clawteam\u002Fteams\u002F\u003Cteam>\u002Finboxes\u002F\u003Cagent>\u002Fmsg-*.json`\n- Costs: `~\u002F.clawteam\u002Fcosts\u002F\u003Cteam>\u002F`\n","# ClawTeam — Multi-Agent Swarm Coordination\n\n## Overview\n\nClawTeam is a CLI tool (`clawteam`) for orchestrating multiple AI agents as self-organizing swarms. It uses git worktree isolation, tmux windows, and filesystem-based messaging. OpenClaw is the default agent backend.\n\n**CLI binary**: `clawteam` (installed via pip, available in PATH)\n\n## Quick Start\n\n### One-Command Template Launch (Recommended)\n\n```bash\n# Launch a pre-built team from a template\nclawteam launch hedge-fund --team fund1\nclawteam launch code-review --team review1\nclawteam launch research-paper --team paper1\n```\n\n### Manual Team Setup\n\n```bash\n# 1. Create team with leader\nclawteam team spawn-team my-team -d \"Build a web app\" -n leader\n\n# 2. Create tasks with dependencies\nclawteam task create my-team \"Design API schema\" -o architect\n# Returns task ID, e.g., abc123\n\nclawteam task create my-team \"Implement auth\" -o backend --blocked-by abc123\nclawteam task create my-team \"Build frontend\" -o frontend --blocked-by abc123\nclawteam task create my-team \"Write tests\" -o tester\n\n# 3. Spawn agents (each gets its own tmux window + git worktree)\nclawteam spawn -t my-team -n architect --task \"Design the API schema for a web app\"\nclawteam spawn -t my-team -n backend --task \"Implement OAuth2 authentication\"\nclawteam spawn -t my-team -n frontend --task \"Build React dashboard\"\n\n# 4. Monitor\nclawteam board show my-team        # Kanban view\nclawteam board attach my-team      # Tmux tiled view (all agents side-by-side)\nclawteam board serve --port 8080   # Web dashboard\n```\n\n## Command Reference\n\n### Team Management\n\n| Command | Description |\n|---------|-------------|\n| `clawteam team spawn-team \u003Cname> -d \"\u003Cdesc>\" -n \u003Cleader>` | Create team |\n| `clawteam team discover` | List all teams |\n| `clawteam team status \u003Cteam>` | Show team members and info |\n| `clawteam team cleanup \u003Cteam> --force` | Delete team and all data |\n\n### Task Management\n\n| Command | Description |\n|---------|-------------|\n| `clawteam task create \u003Cteam> \"\u003Csubject>\" -o \u003Cowner> [-d \"\u003Cdesc>\"] [--blocked-by \u003Cid>]` | Create task |\n| `clawteam task list \u003Cteam> [--owner \u003Cname>]` | List tasks (filterable) |\n| `clawteam task update \u003Cteam> \u003Cid> --status \u003Cstatus>` | Update status |\n| `clawteam task get \u003Cteam> \u003Cid>` | Get single task |\n| `clawteam task stats \u003Cteam>` | Timing statistics |\n| `clawteam task wait \u003Cteam>` | Block until all tasks complete |\n\n**Task statuses**: `pending`, `in_progress`, `completed`, `blocked`\n\n**Dependency auto-resolution**: When a blocking task completes, dependent tasks automatically change from `blocked` to `pending`.\n\n**Task locking**: When a task moves to `in_progress`, it is locked by the calling agent. Other agents cannot claim it unless they use `--force`. Stale locks from dead agents are automatically released.\n\n### Agent Spawning\n\n**IMPORTANT**: Always use the default command (`openclaw`) — do NOT override to `claude` or other agents. The default handles permissions, prompt injection, and nesting detection correctly. If you specify `claude` as the command, agents will get stuck on interactive permission prompts.\n\n```bash\n# Default (RECOMMENDED): spawns openclaw tui in tmux with prompt\nclawteam spawn -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask description>\"\n\n# Explicit backend (still uses openclaw by default)\nclawteam spawn tmux -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask>\"\nclawteam spawn subprocess -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask>\"\n\n# With git worktree isolation\nclawteam spawn -t \u003Cteam> -n \u003Cname> --task \"\u003Ctask>\" --workspace --repo \u002Fpath\u002Fto\u002Frepo\n```\n\nEach spawned agent gets:\n- Its own tmux window (visible via `board attach`)\n- Its own git worktree branch (`clawteam\u002F{team}\u002F{agent}`)\n- An auto-injected coordination prompt (how to use clawteam CLI)\n- Environment variables: `CLAWTEAM_AGENT_NAME`, `CLAWTEAM_TEAM_NAME`, etc.\n\n**Spawn safety features:**\n- Commands are pre-validated before launch — you get a clear error if the agent CLI is not installed\n- If a spawn fails, the registered team member and worktree are automatically rolled back\n- Claude Code and Codex workspace trust prompts are auto-confirmed in fresh worktrees\n\n### Messaging\n\n| Command | Description |\n|---------|-------------|\n| `clawteam inbox send \u003Cteam> \u003Cto> \"\u003Cmsg>\" --from \u003Csender>` | Point-to-point message |\n| `clawteam inbox broadcast \u003Cteam> \"\u003Cmsg>\" --from \u003Csender>` | Broadcast to all |\n| `clawteam inbox peek \u003Cteam> -a \u003Cagent>` | Peek without consuming |\n| `clawteam inbox receive \u003Cteam>` | Consume messages |\n| `clawteam inbox log \u003Cteam>` | View message history |\n\n### Monitoring\n\n| Command | Description |\n|---------|-------------|\n| `clawteam board show \u003Cteam>` | Kanban board (rich terminal) |\n| `clawteam board overview` | All teams overview |\n| `clawteam board live \u003Cteam>` | Live-refreshing board |\n| `clawteam board attach \u003Cteam>` | Tmux tiled view |\n| `clawteam board serve --port 8080` | Web dashboard |\n\n### Cost Tracking\n\n| Command | Description |\n|---------|-------------|\n| `clawteam cost report \u003Cteam> --input-tokens \u003CN> --output-tokens \u003CN> --cost-cents \u003CN>` | Report usage |\n| `clawteam cost show \u003Cteam>` | Show summary |\n| `clawteam cost budget \u003Cteam> \u003Cdollars>` | Set budget |\n\n### Templates\n\n| Command | Description |\n|---------|-------------|\n| `clawteam template list` | List available templates |\n| `clawteam template show \u003Cname>` | Show template details |\n| `clawteam launch \u003Ctemplate> [--team-name \u003Cname>] [--goal \"\u003Cgoal>\"]` | Launch from template |\n\n**Built-in templates**: `hedge-fund`, `code-review`, `research-paper`\n\n### Configuration\n\n```bash\nclawteam config show                           # Show all settings\nclawteam config set transport file             # Set transport backend\nclawteam config set skip_permissions true      # Auto-skip permission prompts\nclawteam config health                         # System health check\n```\n\n### Other Commands\n\n| Command | Description |\n|---------|-------------|\n| `clawteam lifecycle idle \u003Cteam> --agent \u003Cname>` | Report agent idle |\n| `clawteam session save \u003Cteam> --session-id \u003Cid>` | Save session for resume |\n| `clawteam plan submit \u003Cteam> \"\u003Cplan>\" --from \u003Cagent>` | Submit plan for approval (team-scoped storage) |\n| `clawteam workspace list \u003Cteam>` | List git worktrees |\n| `clawteam workspace merge \u003Cteam> --agent \u003Cname>` | Merge agent branch |\n\n## JSON Output\n\nAdd `--json` before any subcommand for machine-readable output:\n\n```bash\nclawteam --json task list my-team\nclawteam --json team status my-team\n```\n\n## Typical Workflow\n\n1. **User says**: \"Create a team to build a web app\"\n2. **You do**: `clawteam team spawn-team webapp -d \"Build web app\" -n leader`\n3. **Create tasks**: Use `clawteam task create` with `--blocked-by` for dependencies\n4. **Spawn agents**: Use `clawteam spawn` for each worker\n5. **Monitor**: Start a background polling loop immediately — do NOT wait for user to ask\n6. **Communicate**: Use `clawteam inbox broadcast` for team-wide updates\n7. **Deliver**: Proactively send final results to the user as soon as all tasks complete\n8. **Cleanup**: `clawteam cost show`, `clawteam task stats`, merge worktrees, then `clawteam team cleanup webapp --force`\n\n## Leader Orchestration Pattern\n\nWhen YOU are the leader agent, follow this pattern to autonomously manage a swarm:\n\n### Phase 1: Analyze & Plan\n```\n1. Understand the user's goal\n2. Break it into independent subtasks\n3. Identify dependencies between tasks (what must finish before what)\n4. Decide how many worker agents are needed\n```\n\n### Phase 2: Setup\n```bash\n# Create team\nclawteam team spawn-team \u003Cteam> -d \"\u003Cgoal description>\" -n leader\n\n# Create tasks with dependency chains\nclawteam task create \u003Cteam> \"Design API\" -o architect\n# Save the returned task ID (e.g., abc123)\nclawteam task create \u003Cteam> \"Build backend\" -o backend --blocked-by abc123\nclawteam task create \u003Cteam> \"Build frontend\" -o frontend --blocked-by abc123\nclawteam task create \u003Cteam> \"Integration tests\" -o tester --blocked-by \u003Cbackend-id>,\u003Cfrontend-id>\n```\n\n### Phase 3: Spawn Workers\n```bash\n# Each spawn launches an openclaw tui in its own tmux window\nclawteam spawn -t \u003Cteam> -n architect --task \"Design REST API schema for \u003Cgoal>\"\nclawteam spawn -t \u003Cteam> -n backend --task \"Implement backend based on API schema\"\nclawteam spawn -t \u003Cteam> -n frontend --task \"Build React frontend\"\nclawteam spawn -t \u003Cteam> -n tester --task \"Write and run integration tests\"\n```\n\n### Phase 4: Monitor Loop\n\n**IMPORTANT**: Start monitoring immediately after spawning — do NOT wait for the user to ask for status updates. Run the monitor loop in the background right away so you can:\n1. **Push mid-progress updates proactively** — when ~50% of tasks complete, send the user a brief status update (e.g. \"4\u002F7 agents done, 3 still working\"). Do NOT wait for them to ask.\n2. **Deliver final results immediately** when all tasks complete.\n\n```bash\n# Poll task status every 30-60 seconds\nwhile true; do\n  clawteam --json task list \u003Cteam> | python3 -c \"\nimport sys, json\ntasks = json.load(sys.stdin)\ndone = sum(1 for t in tasks if t['status'] == 'completed')\ntotal = len(tasks)\nprint(f'{done}\u002F{total} complete')\nif done == total: print('ALL DONE'); sys.exit(0)\n\"\n  # Check for messages from workers\n  clawteam inbox receive \u003Cteam>\n  # IMPORTANT: Send a mid-progress update to the user when roughly half the tasks are done\n  sleep 30\ndone\n```\n\n### Phase 5: Converge & Report\n\n**IMPORTANT**: Proactively deliver results to the user as soon as all tasks complete. Do NOT wait for the user to ask. Include the final output, a summary, and cost\u002Ftiming stats. ALWAYS merge worktrees and clean up.\n\n```bash\n# After all tasks complete — do ALL of these steps:\nclawteam board show \u003Cteam>           # Final status\nclawteam cost show \u003Cteam>            # Total cost — include in report to user\nclawteam task stats \u003Cteam>           # Timing stats — include in report to user\n# Merge each worker's branch back to main\nfor agent in \u003Cagent1> \u003Cagent2> ...; do\n  clawteam workspace merge \u003Cteam> --agent $agent\ndone\nclawteam team cleanup \u003Cteam> --force  # Clean up — ALWAYS do this last\n# Then: send the final deliverables to the user immediately\n```\n\n### Decision Rules for the Leader\n- **Independent tasks** → spawn workers in parallel\n- **Sequential tasks** → use `--blocked-by` to chain them; ClawTeam auto-unblocks\n- **Worker asks for help** → check inbox, provide guidance via `inbox send`\n- **Worker stuck** → check task status; if `in_progress` too long, send a nudge via `inbox send`\n- **Worker done** → verify result via inbox message, then move to next phase\n- **All done** → merge worktrees, deliver results to user proactively, then cleanup\n- **Always** → start background monitoring immediately after spawn; never wait for user to ask for status\n\n## Data Location\n\nAll state stored in `~\u002F.clawteam\u002F`:\n- Teams: `~\u002F.clawteam\u002Fteams\u002F\u003Cteam>\u002Fconfig.json`\n- Tasks: `~\u002F.clawteam\u002Ftasks\u002F\u003Cteam>\u002Ftask-\u003Cid>.json` (with `fcntl` file locking for concurrent safety)\n- Plans: `~\u002F.clawteam\u002Fplans\u002F\u003Cteam>\u002F\u003Cagent>-\u003Cplan_id>.md` (team-scoped, isolated per team)\n- Messages: `~\u002F.clawteam\u002Fteams\u002F\u003Cteam>\u002Finboxes\u002F\u003Cagent>\u002Fmsg-*.json`\n- Costs: `~\u002F.clawteam\u002Fcosts\u002F\u003Cteam>\u002F`\n",{"name":5,"description":7},"main","d69096e239f2a89aabc2cd078c47664a9d3b1e11","05-技能与代理开发\u002Fopenclaw",3,"https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts","https:\u002F\u002Fgithub.com\u002Flza6","个人vx：Tf00798\r\n个人✈：@TFGY999","2026-10-08 14:29:22",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add lza6\u002FArmor-piercing-system-prompts --skill clawteam",{"tool":38,"command":39},"bunx","bunx skills add lza6\u002FArmor-piercing-system-prompts --skill clawteam",{"tool":41,"command":42},"pnpm","pnpm dlx skills add lza6\u002FArmor-piercing-system-prompts --skill clawteam","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F05-技能与代理开发\u002Fopenclaw\u002FSKILL.md",true,[],[47,54,59,66,73,80],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":19},"01M7FFG485KY83TD6HSC4YPCMC","obsidian-clipper-template-creator","Guide for creating templates for the Obsidian Web Clipper. Use when you want to create a new clipping template, understand available variables, or format clipped content.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator\u002FSKILL.md",[],[],{"skillNo":55,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M7FFG0T6AZ0STEJPSB3CX84K","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobsidian-clipper-template-creator-antigravity-awesome-skills-main\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M7FFFY0QEM14SKWPDYGMXM2P","python-api","Programmatic Object-Detection-Metrics API guidance for BoundingBox, BoundingBoxes, Evaluator, enums, coordinate conversion, AP\u002FmAP semantics, and source-style import diagnostics.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Fpython-api\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M7FFFVZQZ2GJ0C0TGP7JHD20","file-evaluation","Folder-based PASCAL VOC AP\u002FmAP evaluation from ground-truth and detection text files.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002Fsub-skills\u002Ffile-evaluation\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M7FFFSFNYSE77SJ6PE1CK2DA","object-detection-metrics","Guides legacy Object-Detection-Metrics PASCAL VOC AP\u002FmAP evaluation from detection text folders and direct Python API classes.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fobject-detection-metrics\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M7FFFP0GCK1RVXBCPQ8APBMT","nx-workspace-patterns","Configure and optimize Nx monorepo workspaces. Use when setting up Nx, configuring project boundaries, optimizing build caching, or implementing affected commands.","https:\u002F\u002Fgithub.com\u002Flza6\u002FArmor-piercing-system-prompts\u002Fblob\u002Fmain\u002F07-后端与架构\u002Fnx-workspace-patterns\u002FSKILL.md",[],[]]